rocksdb/tools/ldb_cmd.cc

4452 lines
153 KiB
C++
Raw Normal View History

// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
// This source code is licensed under both the GPLv2 (found in the
// COPYING file in the root directory) and Apache 2.0 License
// (found in the LICENSE.Apache file in the root directory).
//
#include "rocksdb/utilities/ldb_cmd.h"
GetEntity and PutEntity Support in ldb (#11796) Summary: - `get_entity` and `put_entity` command support in ldb - Input Format for `put_entity`: `ldb --db=<DB_PATH> put_entity <KEY> <COLUMN_1_NAME>:<COLUMN_1_VALUE> <COLUMN_2_NAME>:<COLUMN_2_VALUE> ...` - Output Format for `get_entity`: `<COLUMN_1_NAME>:<COLUMN_1_VALUE> <COLUMN_2_NAME>:<COLUMN_2_VALUE>` - If `get_entity` is called against non-wide column value (existing behavior), empty key (kDefaultWideColumnName) will be printed, appended by `:<COLUMN_VALUE>` - If `get` is called against wide column value (existing behavior), first column value is printed if the first column name is kDefaultWideColumnName. # Test Checks for `put_entity` and `get_entity` added in `ldb_test.py` ``` ❯ python3 tools/ldb_test.py took 45s at 10:45:44 AM Running testBlobBatchPut... .Running testBlobDump .Running testBlobPut... .Running testBlobStartingLevel... .Running testCheckConsistency... .Running testColumnFamilies... .Running testCountDelimDump... .Running testCountDelimIDump... .Running testDumpLiveFiles... .Running testDumpLoad... Warning: 7 bad lines ignored. .Running testGetProperty... .Running testHexPutGet... .Running testIDumpBasics... .Running testIDumpDecodeBlobIndex... .Running testIngestExternalSst... .Running testInvalidCmdLines... .Running testListColumnFamilies... .Running testListLiveFilesMetadata... .Running testManifestDump... .Running testMiscAdminTask... Compacting the db... Sequence,Count,ByteSize,Physical Offset,Key(s) .Running testSSTDump... .Running testSimpleStringPutGet... .Running testStringBatchPut... .Running testTtlPutGet... .Running testWALDump... . ---------------------------------------------------------------------- Ran 25 tests in 57.742s ``` Manual Test ``` # Invalid format for wide columns ❯ ./ldb --db=/tmp/test_db put_entity x4 x5 Failed: wide column format needs to be <column_name>:<column_value> (did you mean put <key> <value>?) # empty column name (kDefaultWideColumnName) ❯ ./ldb --db=/tmp/test_db put_entity x4 :x5 OK ❯ ./ldb --db=/tmp/test_db get_entity x4 :x5 ❯ ./ldb --db=/tmp/test_db get x4 x5 ❯ ./ldb --db=/tmp/test_db put_entity a1 :z1 b1:c1 b2:f1 OK ❯ ./ldb --db=/tmp/test_db get_entity a1 :z1 b1:c1 b2:f1 # Keeping the existing behavior if `get` was called on wide column values ❯ ./ldb --db=/tmp/test_db get a1 z1 # Scan ❯ ./ldb --db=/tmp/test_db scan a1 ==> b1:c1 b2:f1 x4 ==> x5 x5 ==> cn1:cv1 cn2:cv2 # Scan hex ❯ ./ldb --db=/tmp/test_db scan --hex 0x6131 ==> 0x6231:0x6331 0x6232:0x6631 0x7834 ==> 0x7835 0x7835 ==> 0x636E31:0x637631 0x636E32:0x637632 # More testing with hex values ❯ ./ldb --db=/tmp/test_db get_entity 0x6131 --hex 0x6231:0x6331 0x6232:0x6631 ❯ ./ldb --db=/tmp/test_db get_entity 0x78 --hex Failed: GetEntity failed: NotFound: ❯ ./ldb --db=/tmp/test_db get_entity 0x7834 --hex :0x7835 ❯ ./ldb --db=/tmp/test_db put_entity 0x7834 0x6234:0x6635 --hex OK ❯ ./ldb --db=/tmp/test_db get_entity 0x7834 --hex 0x6234:0x6635 ❯ ./ldb --db=/tmp/test_db get_entity 0x7834 --key_hex b4:f5 ❯ ./ldb --db=/tmp/test_db get_entity x4 --value_hex 0x6234:0x6635 ``` Pull Request resolved: https://github.com/facebook/rocksdb/pull/11796 Reviewed By: jowlyzhang Differential Revision: D48978141 Pulled By: jaykorean fbshipit-source-id: 4f87c222417ed90a6dbf39bd7b0f068b01e68393
2023-09-12 23:32:40 +00:00
#include <cstddef>
#include <cstdlib>
#include <ctime>
#include <fstream>
#include <functional>
#include <iostream>
#include <limits>
#include <sstream>
#include <stdexcept>
#include <string>
#include "db/blob/blob_index.h"
#include "db/db_impl/db_impl.h"
#include "db/dbformat.h"
#include "db/log_reader.h"
Add manifest fix-up utility for file temperatures (#9683) Summary: The goal of this change is to allow changes to the "current" (in FileSystem) file temperatures to feed back into DB metadata, so that they can inform decisions and stats reporting. In part because of modular code factoring, it doesn't seem easy to do this automagically, where opening an SST file and observing current Temperature different from expected would trigger a change in metadata and DB manifest write (essentially giving the deep read path access to the write path). It is also difficult to do this while the DB is open because of the limitations of LogAndApply. This change allows updating file temperature metadata on a closed DB using an experimental utility function UpdateManifestForFilesState() or `ldb update_manifest --update_temperatures`. This should suffice for "migration" scenarios where outside tooling has placed or re-arranged DB files into a (different) tiered configuration without going through RocksDB itself (currently, only compaction can change temperature metadata). Some details: * Refactored and added unit test for `ldb unsafe_remove_sst_file` because of shared functionality * Pulled in autovector.h changes from https://github.com/facebook/rocksdb/issues/9546 to fix SuperVersionContext move constructor (related to an older draft of this change) Possible follow-up work: * Support updating manifest with file checksums, such as when a new checksum function is used and want existing DB metadata updated for it. * It's possible that for some repair scenarios, lighter weight than full repair, we might want to support UpdateManifestForFilesState() to modify critical file details like size or checksum using same algorithm. But let's make sure these are differentiated from modifying file details in ways that don't suspect corruption (or require extreme trust). Pull Request resolved: https://github.com/facebook/rocksdb/pull/9683 Test Plan: unit tests added Reviewed By: jay-zhuang Differential Revision: D34798828 Pulled By: pdillinger fbshipit-source-id: cfd83e8fb10761d8c9e7f9c020d68c9106a95554
2022-03-18 23:35:51 +00:00
#include "db/version_util.h"
Wide Column support in ldb (#11754) Summary: wide_columns can now be pretty-printed in the following commands - `./ldb dump_wal` - `./ldb dump` - `./ldb idump` - `./ldb dump_live_files` - `./ldb scan` - `./sst_dump --command=scan` There are opportunities to refactor to reduce some nearly identical code. This PR is initial change to add wide column support in `ldb` and `sst_dump` tool. More PRs to come for the refactor. Pull Request resolved: https://github.com/facebook/rocksdb/pull/11754 Test Plan: **New Tests added** - `WideColumnsHelperTest::DumpWideColumns` - `WideColumnsHelperTest::DumpSliceAsWideColumns` **Changes added to existing tests** - `ExternalSSTFileTest::BasicMixed` added to cover mixed case (This test should have been added in https://github.com/facebook/rocksdb/issues/11688). This test does not verify the ldb or sst_dump output. This test was used to create test SST files having some rows with wide columns and some without and the generated SST files were used to manually test sst_dump_tool. - `createSST()` in `sst_dump_test` now takes `wide_column_one_in` to add wide column value in SST **dump_wal** ``` ./ldb dump_wal --walfile=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/000004.log --print_value --header ``` ``` Sequence,Count,ByteSize,Physical Offset,Key(s) : value 1,1,59,0,PUT_ENTITY(0) : 0x:0x68656C6C6F 0x617474725F6E616D6531:0x666F6F 0x617474725F6E616D6532:0x626172 2,1,34,42,PUT_ENTITY(0) : 0x617474725F6F6E65:0x74776F 0x617474725F7468726565:0x666F7572 3,1,17,7d,PUT(0) : 0x7468697264 : 0x62617A ``` **idump** ``` ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/ idump ``` ``` 'first' seq:1, type:22 => :hello attr_name1:foo attr_name2:bar 'second' seq:2, type:22 => attr_one:two attr_three:four 'third' seq:3, type:1 => baz Internal keys in range: 3 ``` **SST Dump from dump_live_files** ``` ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/ compact ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/ dump_live_files ``` ``` ... ============================== SST Files ============================== /tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/000013.sst level:1 ------------------------------ Process /tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/000013.sst Sst file format: block-based 'first' seq:0, type:22 => :hello attr_name1:foo attr_name2:bar 'second' seq:0, type:22 => attr_one:two attr_three:four 'third' seq:0, type:1 => baz ... ``` **dump** ``` ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/ dump ``` ``` first ==> :hello attr_name1:foo attr_name2:bar second ==> attr_one:two attr_three:four third ==> baz Keys in range: 3 ``` **scan** ``` ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/ scan ``` ``` first : :hello attr_name1:foo attr_name2:bar second : attr_one:two attr_three:four third : baz ``` **sst_dump** ``` ./sst_dump --file=/tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/000013.sst --command=scan ``` ``` options.env is 0x7ff54b296000 Process /tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/000013.sst Sst file format: block-based from [] to [] 'first' seq:0, type:22 => :hello attr_name1:foo attr_name2:bar 'second' seq:0, type:22 => attr_one:two attr_three:four 'third' seq:0, type:1 => baz ``` Reviewed By: ltamasi Differential Revision: D48837999 Pulled By: jaykorean fbshipit-source-id: b0280f0589d2b9716bb9b50530ffcabb397d140f
2023-08-30 19:45:52 +00:00
#include "db/wide/wide_column_serialization.h"
#include "db/wide/wide_columns_helper.h"
#include "db/write_batch_internal.h"
#include "file/filename.h"
#include "rocksdb/cache.h"
Add manifest fix-up utility for file temperatures (#9683) Summary: The goal of this change is to allow changes to the "current" (in FileSystem) file temperatures to feed back into DB metadata, so that they can inform decisions and stats reporting. In part because of modular code factoring, it doesn't seem easy to do this automagically, where opening an SST file and observing current Temperature different from expected would trigger a change in metadata and DB manifest write (essentially giving the deep read path access to the write path). It is also difficult to do this while the DB is open because of the limitations of LogAndApply. This change allows updating file temperature metadata on a closed DB using an experimental utility function UpdateManifestForFilesState() or `ldb update_manifest --update_temperatures`. This should suffice for "migration" scenarios where outside tooling has placed or re-arranged DB files into a (different) tiered configuration without going through RocksDB itself (currently, only compaction can change temperature metadata). Some details: * Refactored and added unit test for `ldb unsafe_remove_sst_file` because of shared functionality * Pulled in autovector.h changes from https://github.com/facebook/rocksdb/issues/9546 to fix SuperVersionContext move constructor (related to an older draft of this change) Possible follow-up work: * Support updating manifest with file checksums, such as when a new checksum function is used and want existing DB metadata updated for it. * It's possible that for some repair scenarios, lighter weight than full repair, we might want to support UpdateManifestForFilesState() to modify critical file details like size or checksum using same algorithm. But let's make sure these are differentiated from modifying file details in ways that don't suspect corruption (or require extreme trust). Pull Request resolved: https://github.com/facebook/rocksdb/pull/9683 Test Plan: unit tests added Reviewed By: jay-zhuang Differential Revision: D34798828 Pulled By: pdillinger fbshipit-source-id: cfd83e8fb10761d8c9e7f9c020d68c9106a95554
2022-03-18 23:35:51 +00:00
#include "rocksdb/experimental.h"
#include "rocksdb/file_checksum.h"
#include "rocksdb/filter_policy.h"
Add manifest fix-up utility for file temperatures (#9683) Summary: The goal of this change is to allow changes to the "current" (in FileSystem) file temperatures to feed back into DB metadata, so that they can inform decisions and stats reporting. In part because of modular code factoring, it doesn't seem easy to do this automagically, where opening an SST file and observing current Temperature different from expected would trigger a change in metadata and DB manifest write (essentially giving the deep read path access to the write path). It is also difficult to do this while the DB is open because of the limitations of LogAndApply. This change allows updating file temperature metadata on a closed DB using an experimental utility function UpdateManifestForFilesState() or `ldb update_manifest --update_temperatures`. This should suffice for "migration" scenarios where outside tooling has placed or re-arranged DB files into a (different) tiered configuration without going through RocksDB itself (currently, only compaction can change temperature metadata). Some details: * Refactored and added unit test for `ldb unsafe_remove_sst_file` because of shared functionality * Pulled in autovector.h changes from https://github.com/facebook/rocksdb/issues/9546 to fix SuperVersionContext move constructor (related to an older draft of this change) Possible follow-up work: * Support updating manifest with file checksums, such as when a new checksum function is used and want existing DB metadata updated for it. * It's possible that for some repair scenarios, lighter weight than full repair, we might want to support UpdateManifestForFilesState() to modify critical file details like size or checksum using same algorithm. But let's make sure these are differentiated from modifying file details in ways that don't suspect corruption (or require extreme trust). Pull Request resolved: https://github.com/facebook/rocksdb/pull/9683 Test Plan: unit tests added Reviewed By: jay-zhuang Differential Revision: D34798828 Pulled By: pdillinger fbshipit-source-id: cfd83e8fb10761d8c9e7f9c020d68c9106a95554
2022-03-18 23:35:51 +00:00
#include "rocksdb/options.h"
#include "rocksdb/table_properties.h"
#include "rocksdb/utilities/backup_engine.h"
#include "rocksdb/utilities/checkpoint.h"
#include "rocksdb/utilities/debug.h"
#include "rocksdb/utilities/options_util.h"
#include "rocksdb/write_batch.h"
#include "rocksdb/write_buffer_manager.h"
#include "table/scoped_arena_iterator.h"
#include "table/sst_file_dumper.h"
#include "tools/ldb_cmd_impl.h"
#include "util/cast_util.h"
#include "util/coding.h"
#include "util/file_checksum_helper.h"
#include "util/stderr_logger.h"
#include "util/string_util.h"
#include "utilities/blob_db/blob_dump_tool.h"
#include "utilities/merge_operators.h"
#include "utilities/ttl/db_ttl_impl.h"
namespace ROCKSDB_NAMESPACE {
class FileChecksumGenCrc32c;
class FileChecksumGenCrc32cFactory;
const std::string LDBCommand::ARG_ENV_URI = "env_uri";
const std::string LDBCommand::ARG_FS_URI = "fs_uri";
const std::string LDBCommand::ARG_DB = "db";
const std::string LDBCommand::ARG_PATH = "path";
const std::string LDBCommand::ARG_SECONDARY_PATH = "secondary_path";
const std::string LDBCommand::ARG_HEX = "hex";
const std::string LDBCommand::ARG_KEY_HEX = "key_hex";
const std::string LDBCommand::ARG_VALUE_HEX = "value_hex";
const std::string LDBCommand::ARG_CF_NAME = "column_family";
const std::string LDBCommand::ARG_TTL = "ttl";
const std::string LDBCommand::ARG_TTL_START = "start_time";
const std::string LDBCommand::ARG_TTL_END = "end_time";
const std::string LDBCommand::ARG_TIMESTAMP = "timestamp";
const std::string LDBCommand::ARG_TRY_LOAD_OPTIONS = "try_load_options";
const std::string LDBCommand::ARG_DISABLE_CONSISTENCY_CHECKS =
"disable_consistency_checks";
const std::string LDBCommand::ARG_IGNORE_UNKNOWN_OPTIONS =
"ignore_unknown_options";
const std::string LDBCommand::ARG_FROM = "from";
const std::string LDBCommand::ARG_TO = "to";
const std::string LDBCommand::ARG_MAX_KEYS = "max_keys";
const std::string LDBCommand::ARG_BLOOM_BITS = "bloom_bits";
const std::string LDBCommand::ARG_FIX_PREFIX_LEN = "fix_prefix_len";
const std::string LDBCommand::ARG_COMPRESSION_TYPE = "compression_type";
const std::string LDBCommand::ARG_COMPRESSION_MAX_DICT_BYTES =
"compression_max_dict_bytes";
const std::string LDBCommand::ARG_BLOCK_SIZE = "block_size";
const std::string LDBCommand::ARG_AUTO_COMPACTION = "auto_compaction";
const std::string LDBCommand::ARG_DB_WRITE_BUFFER_SIZE = "db_write_buffer_size";
const std::string LDBCommand::ARG_WRITE_BUFFER_SIZE = "write_buffer_size";
const std::string LDBCommand::ARG_FILE_SIZE = "file_size";
const std::string LDBCommand::ARG_CREATE_IF_MISSING = "create_if_missing";
const std::string LDBCommand::ARG_NO_VALUE = "no_value";
const std::string LDBCommand::ARG_ENABLE_BLOB_FILES = "enable_blob_files";
const std::string LDBCommand::ARG_MIN_BLOB_SIZE = "min_blob_size";
const std::string LDBCommand::ARG_BLOB_FILE_SIZE = "blob_file_size";
const std::string LDBCommand::ARG_BLOB_COMPRESSION_TYPE =
"blob_compression_type";
const std::string LDBCommand::ARG_ENABLE_BLOB_GARBAGE_COLLECTION =
"enable_blob_garbage_collection";
const std::string LDBCommand::ARG_BLOB_GARBAGE_COLLECTION_AGE_CUTOFF =
"blob_garbage_collection_age_cutoff";
const std::string LDBCommand::ARG_BLOB_GARBAGE_COLLECTION_FORCE_THRESHOLD =
"blob_garbage_collection_force_threshold";
const std::string LDBCommand::ARG_BLOB_COMPACTION_READAHEAD_SIZE =
"blob_compaction_readahead_size";
Make it possible to enable blob files starting from a certain LSM tree level (#10077) Summary: Currently, if blob files are enabled (i.e. `enable_blob_files` is true), large values are extracted both during flush/recovery (when SST files are written into level 0 of the LSM tree) and during compaction into any LSM tree level. For certain use cases that have a mix of short-lived and long-lived values, it might make sense to support extracting large values only during compactions whose output level is greater than or equal to a specified LSM tree level (e.g. compactions into L1/L2/... or above). This could reduce the space amplification caused by large values that are turned into garbage shortly after being written at the price of some write amplification incurred by long-lived values whose extraction to blob files is delayed. In order to achieve this, we would like to do the following: - Add a new configuration option `blob_file_starting_level` (default: 0) to `AdvancedColumnFamilyOptions` (and `MutableCFOptions` and extend the related logic) - Instantiate `BlobFileBuilder` in `BuildTable` (used during flush and recovery, where the LSM tree level is L0) and `CompactionJob` iff `enable_blob_files` is set and the LSM tree level is `>= blob_file_starting_level` - Add unit tests for the new functionality, and add the new option to our stress tests (`db_stress` and `db_crashtest.py` ) - Add the new option to our benchmarking tool `db_bench` and the BlobDB benchmark script `run_blob_bench.sh` - Add the new option to the `ldb` tool (see https://github.com/facebook/rocksdb/wiki/Administration-and-Data-Access-Tool) - Ideally extend the C and Java bindings with the new option - Update the BlobDB wiki to document the new option. Pull Request resolved: https://github.com/facebook/rocksdb/pull/10077 Reviewed By: ltamasi Differential Revision: D36884156 Pulled By: gangliao fbshipit-source-id: 942bab025f04633edca8564ed64791cb5e31627d
2022-06-03 03:04:33 +00:00
const std::string LDBCommand::ARG_BLOB_FILE_STARTING_LEVEL =
"blob_file_starting_level";
const std::string LDBCommand::ARG_PREPOPULATE_BLOB_CACHE =
"prepopulate_blob_cache";
const std::string LDBCommand::ARG_DECODE_BLOB_INDEX = "decode_blob_index";
const std::string LDBCommand::ARG_DUMP_UNCOMPRESSED_BLOBS =
"dump_uncompressed_blobs";
const char* LDBCommand::DELIM = " ==> ";
namespace {
void DumpWalFile(Options options, std::string wal_file, bool print_header,
bool print_values, bool is_write_committed,
LDBCommandExecuteResult* exec_state);
void DumpSstFile(Options options, std::string filename, bool output_hex,
bool show_properties, bool decode_blob_index,
std::string from_key = "", std::string to_key = "");
void DumpBlobFile(const std::string& filename, bool is_key_hex,
bool is_value_hex, bool dump_uncompressed_blobs);
}; // namespace
LDBCommand* LDBCommand::InitFromCmdLineArgs(
int argc, char const* const* argv, const Options& options,
const LDBOptions& ldb_options,
const std::vector<ColumnFamilyDescriptor>* column_families) {
std::vector<std::string> args;
for (int i = 1; i < argc; i++) {
args.push_back(argv[i]);
}
return InitFromCmdLineArgs(args, options, ldb_options, column_families,
SelectCommand);
}
/**
* Parse the command-line arguments and create the appropriate LDBCommand2
* instance.
* The command line arguments must be in the following format:
* ./ldb --db=PATH_TO_DB [--commonOpt1=commonOpt1Val] ..
* COMMAND <PARAM1> <PARAM2> ... [-cmdSpecificOpt1=cmdSpecificOpt1Val] ..
* This is similar to the command line format used by HBaseClientTool.
* Command name is not included in args.
* Returns nullptr if the command-line cannot be parsed.
*/
LDBCommand* LDBCommand::InitFromCmdLineArgs(
const std::vector<std::string>& args, const Options& options,
const LDBOptions& ldb_options,
const std::vector<ColumnFamilyDescriptor>* /*column_families*/,
const std::function<LDBCommand*(const ParsedParams&)>& selector) {
// --x=y command line arguments are added as x->y map entries in
// parsed_params.option_map.
//
// Command-line arguments of the form --hex end up in this array as hex to
// parsed_params.flags
ParsedParams parsed_params;
// Everything other than option_map and flags. Represents commands
// and their parameters. For eg: put key1 value1 go into this vector.
std::vector<std::string> cmdTokens;
const std::string OPTION_PREFIX = "--";
for (const auto& arg : args) {
if (arg[0] == '-' && arg[1] == '-') {
std::vector<std::string> splits = StringSplit(arg, '=');
// --option_name=option_value
if (splits.size() == 2) {
std::string optionKey = splits[0].substr(OPTION_PREFIX.size());
parsed_params.option_map[optionKey] = splits[1];
} else if (splits.size() == 1) {
// --flag_name
std::string optionKey = splits[0].substr(OPTION_PREFIX.size());
parsed_params.flags.push_back(optionKey);
} else {
// --option_name=option_value, option_value contains '='
std::string optionKey = splits[0].substr(OPTION_PREFIX.size());
parsed_params.option_map[optionKey] =
arg.substr(splits[0].length() + 1);
}
} else {
cmdTokens.push_back(arg);
}
}
if (cmdTokens.size() < 1) {
fprintf(stderr, "Command not specified!");
return nullptr;
}
parsed_params.cmd = cmdTokens[0];
parsed_params.cmd_params.assign(cmdTokens.begin() + 1, cmdTokens.end());
LDBCommand* command = selector(parsed_params);
if (command) {
command->SetDBOptions(options);
command->SetLDBOptions(ldb_options);
}
return command;
}
LDBCommand* LDBCommand::SelectCommand(const ParsedParams& parsed_params) {
if (parsed_params.cmd == GetCommand::Name()) {
return new GetCommand(parsed_params.cmd_params, parsed_params.option_map,
parsed_params.flags);
GetEntity and PutEntity Support in ldb (#11796) Summary: - `get_entity` and `put_entity` command support in ldb - Input Format for `put_entity`: `ldb --db=<DB_PATH> put_entity <KEY> <COLUMN_1_NAME>:<COLUMN_1_VALUE> <COLUMN_2_NAME>:<COLUMN_2_VALUE> ...` - Output Format for `get_entity`: `<COLUMN_1_NAME>:<COLUMN_1_VALUE> <COLUMN_2_NAME>:<COLUMN_2_VALUE>` - If `get_entity` is called against non-wide column value (existing behavior), empty key (kDefaultWideColumnName) will be printed, appended by `:<COLUMN_VALUE>` - If `get` is called against wide column value (existing behavior), first column value is printed if the first column name is kDefaultWideColumnName. # Test Checks for `put_entity` and `get_entity` added in `ldb_test.py` ``` ❯ python3 tools/ldb_test.py took 45s at 10:45:44 AM Running testBlobBatchPut... .Running testBlobDump .Running testBlobPut... .Running testBlobStartingLevel... .Running testCheckConsistency... .Running testColumnFamilies... .Running testCountDelimDump... .Running testCountDelimIDump... .Running testDumpLiveFiles... .Running testDumpLoad... Warning: 7 bad lines ignored. .Running testGetProperty... .Running testHexPutGet... .Running testIDumpBasics... .Running testIDumpDecodeBlobIndex... .Running testIngestExternalSst... .Running testInvalidCmdLines... .Running testListColumnFamilies... .Running testListLiveFilesMetadata... .Running testManifestDump... .Running testMiscAdminTask... Compacting the db... Sequence,Count,ByteSize,Physical Offset,Key(s) .Running testSSTDump... .Running testSimpleStringPutGet... .Running testStringBatchPut... .Running testTtlPutGet... .Running testWALDump... . ---------------------------------------------------------------------- Ran 25 tests in 57.742s ``` Manual Test ``` # Invalid format for wide columns ❯ ./ldb --db=/tmp/test_db put_entity x4 x5 Failed: wide column format needs to be <column_name>:<column_value> (did you mean put <key> <value>?) # empty column name (kDefaultWideColumnName) ❯ ./ldb --db=/tmp/test_db put_entity x4 :x5 OK ❯ ./ldb --db=/tmp/test_db get_entity x4 :x5 ❯ ./ldb --db=/tmp/test_db get x4 x5 ❯ ./ldb --db=/tmp/test_db put_entity a1 :z1 b1:c1 b2:f1 OK ❯ ./ldb --db=/tmp/test_db get_entity a1 :z1 b1:c1 b2:f1 # Keeping the existing behavior if `get` was called on wide column values ❯ ./ldb --db=/tmp/test_db get a1 z1 # Scan ❯ ./ldb --db=/tmp/test_db scan a1 ==> b1:c1 b2:f1 x4 ==> x5 x5 ==> cn1:cv1 cn2:cv2 # Scan hex ❯ ./ldb --db=/tmp/test_db scan --hex 0x6131 ==> 0x6231:0x6331 0x6232:0x6631 0x7834 ==> 0x7835 0x7835 ==> 0x636E31:0x637631 0x636E32:0x637632 # More testing with hex values ❯ ./ldb --db=/tmp/test_db get_entity 0x6131 --hex 0x6231:0x6331 0x6232:0x6631 ❯ ./ldb --db=/tmp/test_db get_entity 0x78 --hex Failed: GetEntity failed: NotFound: ❯ ./ldb --db=/tmp/test_db get_entity 0x7834 --hex :0x7835 ❯ ./ldb --db=/tmp/test_db put_entity 0x7834 0x6234:0x6635 --hex OK ❯ ./ldb --db=/tmp/test_db get_entity 0x7834 --hex 0x6234:0x6635 ❯ ./ldb --db=/tmp/test_db get_entity 0x7834 --key_hex b4:f5 ❯ ./ldb --db=/tmp/test_db get_entity x4 --value_hex 0x6234:0x6635 ``` Pull Request resolved: https://github.com/facebook/rocksdb/pull/11796 Reviewed By: jowlyzhang Differential Revision: D48978141 Pulled By: jaykorean fbshipit-source-id: 4f87c222417ed90a6dbf39bd7b0f068b01e68393
2023-09-12 23:32:40 +00:00
} else if (parsed_params.cmd == GetEntityCommand::Name()) {
return new GetEntityCommand(parsed_params.cmd_params,
parsed_params.option_map, parsed_params.flags);
} else if (parsed_params.cmd == PutCommand::Name()) {
return new PutCommand(parsed_params.cmd_params, parsed_params.option_map,
parsed_params.flags);
GetEntity and PutEntity Support in ldb (#11796) Summary: - `get_entity` and `put_entity` command support in ldb - Input Format for `put_entity`: `ldb --db=<DB_PATH> put_entity <KEY> <COLUMN_1_NAME>:<COLUMN_1_VALUE> <COLUMN_2_NAME>:<COLUMN_2_VALUE> ...` - Output Format for `get_entity`: `<COLUMN_1_NAME>:<COLUMN_1_VALUE> <COLUMN_2_NAME>:<COLUMN_2_VALUE>` - If `get_entity` is called against non-wide column value (existing behavior), empty key (kDefaultWideColumnName) will be printed, appended by `:<COLUMN_VALUE>` - If `get` is called against wide column value (existing behavior), first column value is printed if the first column name is kDefaultWideColumnName. # Test Checks for `put_entity` and `get_entity` added in `ldb_test.py` ``` ❯ python3 tools/ldb_test.py took 45s at 10:45:44 AM Running testBlobBatchPut... .Running testBlobDump .Running testBlobPut... .Running testBlobStartingLevel... .Running testCheckConsistency... .Running testColumnFamilies... .Running testCountDelimDump... .Running testCountDelimIDump... .Running testDumpLiveFiles... .Running testDumpLoad... Warning: 7 bad lines ignored. .Running testGetProperty... .Running testHexPutGet... .Running testIDumpBasics... .Running testIDumpDecodeBlobIndex... .Running testIngestExternalSst... .Running testInvalidCmdLines... .Running testListColumnFamilies... .Running testListLiveFilesMetadata... .Running testManifestDump... .Running testMiscAdminTask... Compacting the db... Sequence,Count,ByteSize,Physical Offset,Key(s) .Running testSSTDump... .Running testSimpleStringPutGet... .Running testStringBatchPut... .Running testTtlPutGet... .Running testWALDump... . ---------------------------------------------------------------------- Ran 25 tests in 57.742s ``` Manual Test ``` # Invalid format for wide columns ❯ ./ldb --db=/tmp/test_db put_entity x4 x5 Failed: wide column format needs to be <column_name>:<column_value> (did you mean put <key> <value>?) # empty column name (kDefaultWideColumnName) ❯ ./ldb --db=/tmp/test_db put_entity x4 :x5 OK ❯ ./ldb --db=/tmp/test_db get_entity x4 :x5 ❯ ./ldb --db=/tmp/test_db get x4 x5 ❯ ./ldb --db=/tmp/test_db put_entity a1 :z1 b1:c1 b2:f1 OK ❯ ./ldb --db=/tmp/test_db get_entity a1 :z1 b1:c1 b2:f1 # Keeping the existing behavior if `get` was called on wide column values ❯ ./ldb --db=/tmp/test_db get a1 z1 # Scan ❯ ./ldb --db=/tmp/test_db scan a1 ==> b1:c1 b2:f1 x4 ==> x5 x5 ==> cn1:cv1 cn2:cv2 # Scan hex ❯ ./ldb --db=/tmp/test_db scan --hex 0x6131 ==> 0x6231:0x6331 0x6232:0x6631 0x7834 ==> 0x7835 0x7835 ==> 0x636E31:0x637631 0x636E32:0x637632 # More testing with hex values ❯ ./ldb --db=/tmp/test_db get_entity 0x6131 --hex 0x6231:0x6331 0x6232:0x6631 ❯ ./ldb --db=/tmp/test_db get_entity 0x78 --hex Failed: GetEntity failed: NotFound: ❯ ./ldb --db=/tmp/test_db get_entity 0x7834 --hex :0x7835 ❯ ./ldb --db=/tmp/test_db put_entity 0x7834 0x6234:0x6635 --hex OK ❯ ./ldb --db=/tmp/test_db get_entity 0x7834 --hex 0x6234:0x6635 ❯ ./ldb --db=/tmp/test_db get_entity 0x7834 --key_hex b4:f5 ❯ ./ldb --db=/tmp/test_db get_entity x4 --value_hex 0x6234:0x6635 ``` Pull Request resolved: https://github.com/facebook/rocksdb/pull/11796 Reviewed By: jowlyzhang Differential Revision: D48978141 Pulled By: jaykorean fbshipit-source-id: 4f87c222417ed90a6dbf39bd7b0f068b01e68393
2023-09-12 23:32:40 +00:00
} else if (parsed_params.cmd == PutEntityCommand::Name()) {
return new PutEntityCommand(parsed_params.cmd_params,
parsed_params.option_map, parsed_params.flags);
} else if (parsed_params.cmd == BatchPutCommand::Name()) {
return new BatchPutCommand(parsed_params.cmd_params,
parsed_params.option_map, parsed_params.flags);
} else if (parsed_params.cmd == ScanCommand::Name()) {
return new ScanCommand(parsed_params.cmd_params, parsed_params.option_map,
parsed_params.flags);
} else if (parsed_params.cmd == DeleteCommand::Name()) {
return new DeleteCommand(parsed_params.cmd_params, parsed_params.option_map,
parsed_params.flags);
} else if (parsed_params.cmd == SingleDeleteCommand::Name()) {
return new SingleDeleteCommand(parsed_params.cmd_params,
parsed_params.option_map,
parsed_params.flags);
} else if (parsed_params.cmd == DeleteRangeCommand::Name()) {
return new DeleteRangeCommand(parsed_params.cmd_params,
parsed_params.option_map,
parsed_params.flags);
} else if (parsed_params.cmd == ApproxSizeCommand::Name()) {
return new ApproxSizeCommand(parsed_params.cmd_params,
parsed_params.option_map, parsed_params.flags);
} else if (parsed_params.cmd == DBQuerierCommand::Name()) {
return new DBQuerierCommand(parsed_params.cmd_params,
parsed_params.option_map, parsed_params.flags);
} else if (parsed_params.cmd == CompactorCommand::Name()) {
return new CompactorCommand(parsed_params.cmd_params,
parsed_params.option_map, parsed_params.flags);
} else if (parsed_params.cmd == WALDumperCommand::Name()) {
return new WALDumperCommand(parsed_params.cmd_params,
parsed_params.option_map, parsed_params.flags);
} else if (parsed_params.cmd == ReduceDBLevelsCommand::Name()) {
return new ReduceDBLevelsCommand(parsed_params.cmd_params,
parsed_params.option_map,
parsed_params.flags);
} else if (parsed_params.cmd == ChangeCompactionStyleCommand::Name()) {
return new ChangeCompactionStyleCommand(parsed_params.cmd_params,
parsed_params.option_map,
parsed_params.flags);
} else if (parsed_params.cmd == DBDumperCommand::Name()) {
return new DBDumperCommand(parsed_params.cmd_params,
parsed_params.option_map, parsed_params.flags);
} else if (parsed_params.cmd == DBLoaderCommand::Name()) {
return new DBLoaderCommand(parsed_params.cmd_params,
parsed_params.option_map, parsed_params.flags);
} else if (parsed_params.cmd == ManifestDumpCommand::Name()) {
return new ManifestDumpCommand(parsed_params.cmd_params,
parsed_params.option_map,
parsed_params.flags);
} else if (parsed_params.cmd == FileChecksumDumpCommand::Name()) {
return new FileChecksumDumpCommand(parsed_params.cmd_params,
parsed_params.option_map,
parsed_params.flags);
aggregated-table-properties with GetMapProperty (#7779) Summary: So that we can more easily get aggregate live table data such as total filter, index, and data sizes. Also adds ldb support for getting properties Also fixed some missing/inaccurate related comments in db.h For example: $ ./ldb --db=testdb get_property rocksdb.aggregated-table-properties rocksdb.aggregated-table-properties.data_size: 102871 rocksdb.aggregated-table-properties.filter_size: 0 rocksdb.aggregated-table-properties.index_partitions: 0 rocksdb.aggregated-table-properties.index_size: 2232 rocksdb.aggregated-table-properties.num_data_blocks: 100 rocksdb.aggregated-table-properties.num_deletions: 0 rocksdb.aggregated-table-properties.num_entries: 15000 rocksdb.aggregated-table-properties.num_merge_operands: 0 rocksdb.aggregated-table-properties.num_range_deletions: 0 rocksdb.aggregated-table-properties.raw_key_size: 288890 rocksdb.aggregated-table-properties.raw_value_size: 198890 rocksdb.aggregated-table-properties.top_level_index_size: 0 $ ./ldb --db=testdb get_property rocksdb.aggregated-table-properties-at-level1 rocksdb.aggregated-table-properties-at-level1.data_size: 80909 rocksdb.aggregated-table-properties-at-level1.filter_size: 0 rocksdb.aggregated-table-properties-at-level1.index_partitions: 0 rocksdb.aggregated-table-properties-at-level1.index_size: 1787 rocksdb.aggregated-table-properties-at-level1.num_data_blocks: 81 rocksdb.aggregated-table-properties-at-level1.num_deletions: 0 rocksdb.aggregated-table-properties-at-level1.num_entries: 12466 rocksdb.aggregated-table-properties-at-level1.num_merge_operands: 0 rocksdb.aggregated-table-properties-at-level1.num_range_deletions: 0 rocksdb.aggregated-table-properties-at-level1.raw_key_size: 238210 rocksdb.aggregated-table-properties-at-level1.raw_value_size: 163414 rocksdb.aggregated-table-properties-at-level1.top_level_index_size: 0 $ Pull Request resolved: https://github.com/facebook/rocksdb/pull/7779 Test Plan: Added a test to ldb_test.py Reviewed By: jay-zhuang Differential Revision: D25653103 Pulled By: pdillinger fbshipit-source-id: 2905469a08a64dd6b5510cbd7be2e64d3234d6d3
2020-12-19 15:59:08 +00:00
} else if (parsed_params.cmd == GetPropertyCommand::Name()) {
return new GetPropertyCommand(parsed_params.cmd_params,
parsed_params.option_map,
parsed_params.flags);
} else if (parsed_params.cmd == ListColumnFamiliesCommand::Name()) {
return new ListColumnFamiliesCommand(parsed_params.cmd_params,
parsed_params.option_map,
parsed_params.flags);
} else if (parsed_params.cmd == CreateColumnFamilyCommand::Name()) {
return new CreateColumnFamilyCommand(parsed_params.cmd_params,
parsed_params.option_map,
parsed_params.flags);
} else if (parsed_params.cmd == DropColumnFamilyCommand::Name()) {
return new DropColumnFamilyCommand(parsed_params.cmd_params,
parsed_params.option_map,
parsed_params.flags);
} else if (parsed_params.cmd == DBFileDumperCommand::Name()) {
return new DBFileDumperCommand(parsed_params.cmd_params,
parsed_params.option_map,
parsed_params.flags);
Add list live files metadata (#8446) Summary: Add an argument to ldb to dump live file names, column families, and levels, `list_live_files_metadata`. The output shows all active SST file names, sorted first by column family and then by level. For each level the SST files are sorted alphabetically. Typically, the output looks like this: ``` ./ldb --db=/tmp/test_db list_live_files_metadata Live SST Files: ===== Column Family: default ===== ---------- level 0 ---------- /tmp/test_db/000069.sst ---------- level 1 ---------- /tmp/test_db/000064.sst /tmp/test_db/000065.sst /tmp/test_db/000066.sst /tmp/test_db/000071.sst ---------- level 2 ---------- /tmp/test_db/000038.sst /tmp/test_db/000039.sst /tmp/test_db/000052.sst /tmp/test_db/000067.sst /tmp/test_db/000070.sst ------------------------------ ``` Second, a flag was added `--sort_by_filename`, to change the layout of the output. When this flag is added to the command, the output shows all active SST files sorted by name, in front of which the LSM level and the column family are mentioned. With the same example, the following command would return: ``` ./ldb --db=/tmp/test_db list_live_files_metadata --sort_by_filename Live SST Files: /tmp/test_db/000038.sst : level 2, column family 'default' /tmp/test_db/000039.sst : level 2, column family 'default' /tmp/test_db/000052.sst : level 2, column family 'default' /tmp/test_db/000064.sst : level 1, column family 'default' /tmp/test_db/000065.sst : level 1, column family 'default' /tmp/test_db/000066.sst : level 1, column family 'default' /tmp/test_db/000067.sst : level 2, column family 'default' /tmp/test_db/000069.sst : level 0, column family 'default' /tmp/test_db/000070.sst : level 2, column family 'default' /tmp/test_db/000071.sst : level 1, column family 'default' ------------------------------ ``` Thus, the user can either request to show the files by levels, or sorted by filenames. This PR includes a simple Python unit test that makes sure the file name and level printed out by this new feature matches the one found with an existing feature, `dump_live_file`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8446 Reviewed By: akankshamahajan15 Differential Revision: D29320080 Pulled By: bjlemaire fbshipit-source-id: 01fb7b5637c59010d74c80730a28d815994e7009
2021-06-23 02:06:44 +00:00
} else if (parsed_params.cmd == DBLiveFilesMetadataDumperCommand::Name()) {
return new DBLiveFilesMetadataDumperCommand(parsed_params.cmd_params,
parsed_params.option_map,
parsed_params.flags);
} else if (parsed_params.cmd == InternalDumpCommand::Name()) {
return new InternalDumpCommand(parsed_params.cmd_params,
parsed_params.option_map,
parsed_params.flags);
} else if (parsed_params.cmd == CheckConsistencyCommand::Name()) {
return new CheckConsistencyCommand(parsed_params.cmd_params,
parsed_params.option_map,
parsed_params.flags);
} else if (parsed_params.cmd == CheckPointCommand::Name()) {
return new CheckPointCommand(parsed_params.cmd_params,
parsed_params.option_map, parsed_params.flags);
} else if (parsed_params.cmd == RepairCommand::Name()) {
return new RepairCommand(parsed_params.cmd_params, parsed_params.option_map,
parsed_params.flags);
} else if (parsed_params.cmd == BackupCommand::Name()) {
return new BackupCommand(parsed_params.cmd_params, parsed_params.option_map,
parsed_params.flags);
} else if (parsed_params.cmd == RestoreCommand::Name()) {
return new RestoreCommand(parsed_params.cmd_params,
parsed_params.option_map, parsed_params.flags);
} else if (parsed_params.cmd == WriteExternalSstFilesCommand::Name()) {
return new WriteExternalSstFilesCommand(parsed_params.cmd_params,
parsed_params.option_map,
parsed_params.flags);
} else if (parsed_params.cmd == IngestExternalSstFilesCommand::Name()) {
return new IngestExternalSstFilesCommand(parsed_params.cmd_params,
parsed_params.option_map,
parsed_params.flags);
} else if (parsed_params.cmd == ListFileRangeDeletesCommand::Name()) {
return new ListFileRangeDeletesCommand(parsed_params.option_map,
parsed_params.flags);
} else if (parsed_params.cmd == UnsafeRemoveSstFileCommand::Name()) {
return new UnsafeRemoveSstFileCommand(parsed_params.cmd_params,
parsed_params.option_map,
parsed_params.flags);
Add manifest fix-up utility for file temperatures (#9683) Summary: The goal of this change is to allow changes to the "current" (in FileSystem) file temperatures to feed back into DB metadata, so that they can inform decisions and stats reporting. In part because of modular code factoring, it doesn't seem easy to do this automagically, where opening an SST file and observing current Temperature different from expected would trigger a change in metadata and DB manifest write (essentially giving the deep read path access to the write path). It is also difficult to do this while the DB is open because of the limitations of LogAndApply. This change allows updating file temperature metadata on a closed DB using an experimental utility function UpdateManifestForFilesState() or `ldb update_manifest --update_temperatures`. This should suffice for "migration" scenarios where outside tooling has placed or re-arranged DB files into a (different) tiered configuration without going through RocksDB itself (currently, only compaction can change temperature metadata). Some details: * Refactored and added unit test for `ldb unsafe_remove_sst_file` because of shared functionality * Pulled in autovector.h changes from https://github.com/facebook/rocksdb/issues/9546 to fix SuperVersionContext move constructor (related to an older draft of this change) Possible follow-up work: * Support updating manifest with file checksums, such as when a new checksum function is used and want existing DB metadata updated for it. * It's possible that for some repair scenarios, lighter weight than full repair, we might want to support UpdateManifestForFilesState() to modify critical file details like size or checksum using same algorithm. But let's make sure these are differentiated from modifying file details in ways that don't suspect corruption (or require extreme trust). Pull Request resolved: https://github.com/facebook/rocksdb/pull/9683 Test Plan: unit tests added Reviewed By: jay-zhuang Differential Revision: D34798828 Pulled By: pdillinger fbshipit-source-id: cfd83e8fb10761d8c9e7f9c020d68c9106a95554
2022-03-18 23:35:51 +00:00
} else if (parsed_params.cmd == UpdateManifestCommand::Name()) {
return new UpdateManifestCommand(parsed_params.cmd_params,
parsed_params.option_map,
parsed_params.flags);
}
return nullptr;
}
/* Run the command, and return the execute result. */
void LDBCommand::Run() {
if (!exec_state_.IsNotStarted()) {
return;
}
if (!options_.env || options_.env == Env::Default()) {
Env* env = Env::Default();
Status s = Env::CreateFromUri(config_options_, env_uri_, fs_uri_, &env,
&env_guard_);
if (!s.ok()) {
fprintf(stderr, "%s\n", s.ToString().c_str());
exec_state_ = LDBCommandExecuteResult::Failed(s.ToString());
return;
}
options_.env = env;
}
if (db_ == nullptr && !NoDBOpen()) {
OpenDB();
if (exec_state_.IsFailed() && try_load_options_) {
// We don't always return if there is a failure because a WAL file or
// manifest file can be given to "dump" command so we should continue.
// --try_load_options is not valid in those cases.
return;
}
}
// We'll intentionally proceed even if the DB can't be opened because users
// can also specify a filename, not just a directory.
DoCommand();
if (exec_state_.IsNotStarted()) {
exec_state_ = LDBCommandExecuteResult::Succeed("");
}
if (db_ != nullptr) {
CloseDB();
}
}
LDBCommand::LDBCommand(const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags, bool is_read_only,
const std::vector<std::string>& valid_cmd_line_options)
: db_(nullptr),
db_ttl_(nullptr),
is_read_only_(is_read_only),
is_key_hex_(false),
is_value_hex_(false),
is_db_ttl_(false),
timestamp_(false),
try_load_options_(false),
create_if_missing_(false),
option_map_(options),
flags_(flags),
valid_cmd_line_options_(valid_cmd_line_options) {
auto itr = options.find(ARG_DB);
if (itr != options.end()) {
db_path_ = itr->second;
}
itr = options.find(ARG_ENV_URI);
if (itr != options.end()) {
env_uri_ = itr->second;
}
itr = options.find(ARG_FS_URI);
if (itr != options.end()) {
fs_uri_ = itr->second;
}
itr = options.find(ARG_CF_NAME);
if (itr != options.end()) {
column_family_name_ = itr->second;
} else {
column_family_name_ = kDefaultColumnFamilyName;
}
itr = options.find(ARG_SECONDARY_PATH);
secondary_path_ = "";
if (itr != options.end()) {
secondary_path_ = itr->second;
}
is_key_hex_ = IsKeyHex(options, flags);
is_value_hex_ = IsValueHex(options, flags);
is_db_ttl_ = IsFlagPresent(flags, ARG_TTL);
timestamp_ = IsFlagPresent(flags, ARG_TIMESTAMP);
try_load_options_ = IsTryLoadOptions(options, flags);
force_consistency_checks_ =
!IsFlagPresent(flags, ARG_DISABLE_CONSISTENCY_CHECKS);
enable_blob_files_ = IsFlagPresent(flags, ARG_ENABLE_BLOB_FILES);
enable_blob_garbage_collection_ =
IsFlagPresent(flags, ARG_ENABLE_BLOB_GARBAGE_COLLECTION);
config_options_.ignore_unknown_options =
IsFlagPresent(flags, ARG_IGNORE_UNKNOWN_OPTIONS);
}
void LDBCommand::OpenDB() {
PrepareOptions();
if (!exec_state_.IsNotStarted()) {
return;
}
if (column_families_.empty() && !options_.merge_operator) {
// No harm to add a general merge operator if it is not specified.
options_.merge_operator = MergeOperators::CreateStringAppendOperator(':');
}
// Open the DB.
Status st;
std::vector<ColumnFamilyHandle*> handles_opened;
if (is_db_ttl_) {
// ldb doesn't yet support TTL DB with multiple column families
if (!column_family_name_.empty() || !column_families_.empty()) {
exec_state_ = LDBCommandExecuteResult::Failed(
"ldb doesn't support TTL DB with multiple column families");
}
if (!secondary_path_.empty()) {
exec_state_ = LDBCommandExecuteResult::Failed(
"Open as secondary is not supported for TTL DB yet.");
}
if (is_read_only_) {
st = DBWithTTL::Open(options_, db_path_, &db_ttl_, 0, true);
} else {
st = DBWithTTL::Open(options_, db_path_, &db_ttl_);
}
db_ = db_ttl_;
} else {
if (is_read_only_ && secondary_path_.empty()) {
if (column_families_.empty()) {
st = DB::OpenForReadOnly(options_, db_path_, &db_);
} else {
st = DB::OpenForReadOnly(options_, db_path_, column_families_,
&handles_opened, &db_);
}
} else {
if (column_families_.empty()) {
if (secondary_path_.empty()) {
st = DB::Open(options_, db_path_, &db_);
} else {
st = DB::OpenAsSecondary(options_, db_path_, secondary_path_, &db_);
}
} else {
if (secondary_path_.empty()) {
st = DB::Open(options_, db_path_, column_families_, &handles_opened,
&db_);
} else {
st = DB::OpenAsSecondary(options_, db_path_, secondary_path_,
column_families_, &handles_opened, &db_);
}
}
}
}
if (!st.ok()) {
std::string msg = st.ToString();
exec_state_ = LDBCommandExecuteResult::Failed(msg);
} else if (!handles_opened.empty()) {
assert(handles_opened.size() == column_families_.size());
bool found_cf_name = false;
for (size_t i = 0; i < handles_opened.size(); i++) {
cf_handles_[column_families_[i].name] = handles_opened[i];
if (column_family_name_ == column_families_[i].name) {
found_cf_name = true;
}
}
if (!found_cf_name) {
exec_state_ = LDBCommandExecuteResult::Failed(
"Non-existing column family " + column_family_name_);
CloseDB();
}
} else {
// We successfully opened DB in single column family mode.
assert(column_families_.empty());
if (column_family_name_ != kDefaultColumnFamilyName) {
exec_state_ = LDBCommandExecuteResult::Failed(
"Non-existing column family " + column_family_name_);
CloseDB();
}
}
}
void LDBCommand::CloseDB() {
if (db_ != nullptr) {
for (auto& pair : cf_handles_) {
delete pair.second;
}
Status s = db_->Close();
s.PermitUncheckedError();
delete db_;
db_ = nullptr;
}
}
ColumnFamilyHandle* LDBCommand::GetCfHandle() {
if (!cf_handles_.empty()) {
auto it = cf_handles_.find(column_family_name_);
if (it == cf_handles_.end()) {
exec_state_ = LDBCommandExecuteResult::Failed(
"Cannot find column family " + column_family_name_);
} else {
return it->second;
}
}
return db_->DefaultColumnFamily();
}
std::vector<std::string> LDBCommand::BuildCmdLineOptions(
std::vector<std::string> options) {
std::vector<std::string> ret = {ARG_ENV_URI,
ARG_FS_URI,
ARG_DB,
ARG_SECONDARY_PATH,
ARG_BLOOM_BITS,
ARG_BLOCK_SIZE,
ARG_AUTO_COMPACTION,
ARG_COMPRESSION_TYPE,
ARG_COMPRESSION_MAX_DICT_BYTES,
ARG_WRITE_BUFFER_SIZE,
ARG_FILE_SIZE,
ARG_FIX_PREFIX_LEN,
ARG_TRY_LOAD_OPTIONS,
ARG_DISABLE_CONSISTENCY_CHECKS,
ARG_ENABLE_BLOB_FILES,
ARG_MIN_BLOB_SIZE,
ARG_BLOB_FILE_SIZE,
ARG_BLOB_COMPRESSION_TYPE,
ARG_ENABLE_BLOB_GARBAGE_COLLECTION,
ARG_BLOB_GARBAGE_COLLECTION_AGE_CUTOFF,
ARG_BLOB_GARBAGE_COLLECTION_FORCE_THRESHOLD,
ARG_BLOB_COMPACTION_READAHEAD_SIZE,
Make it possible to enable blob files starting from a certain LSM tree level (#10077) Summary: Currently, if blob files are enabled (i.e. `enable_blob_files` is true), large values are extracted both during flush/recovery (when SST files are written into level 0 of the LSM tree) and during compaction into any LSM tree level. For certain use cases that have a mix of short-lived and long-lived values, it might make sense to support extracting large values only during compactions whose output level is greater than or equal to a specified LSM tree level (e.g. compactions into L1/L2/... or above). This could reduce the space amplification caused by large values that are turned into garbage shortly after being written at the price of some write amplification incurred by long-lived values whose extraction to blob files is delayed. In order to achieve this, we would like to do the following: - Add a new configuration option `blob_file_starting_level` (default: 0) to `AdvancedColumnFamilyOptions` (and `MutableCFOptions` and extend the related logic) - Instantiate `BlobFileBuilder` in `BuildTable` (used during flush and recovery, where the LSM tree level is L0) and `CompactionJob` iff `enable_blob_files` is set and the LSM tree level is `>= blob_file_starting_level` - Add unit tests for the new functionality, and add the new option to our stress tests (`db_stress` and `db_crashtest.py` ) - Add the new option to our benchmarking tool `db_bench` and the BlobDB benchmark script `run_blob_bench.sh` - Add the new option to the `ldb` tool (see https://github.com/facebook/rocksdb/wiki/Administration-and-Data-Access-Tool) - Ideally extend the C and Java bindings with the new option - Update the BlobDB wiki to document the new option. Pull Request resolved: https://github.com/facebook/rocksdb/pull/10077 Reviewed By: ltamasi Differential Revision: D36884156 Pulled By: gangliao fbshipit-source-id: 942bab025f04633edca8564ed64791cb5e31627d
2022-06-03 03:04:33 +00:00
ARG_BLOB_FILE_STARTING_LEVEL,
ARG_PREPOPULATE_BLOB_CACHE,
ARG_IGNORE_UNKNOWN_OPTIONS,
ARG_CF_NAME};
ret.insert(ret.end(), options.begin(), options.end());
return ret;
}
/**
* Parses the specific double option and fills in the value.
* Returns true if the option is found.
* Returns false if the option is not found or if there is an error parsing the
* value. If there is an error, the specified exec_state is also
* updated.
*/
bool LDBCommand::ParseDoubleOption(
const std::map<std::string, std::string>& /*options*/,
const std::string& option, double& value,
LDBCommandExecuteResult& exec_state) {
auto itr = option_map_.find(option);
if (itr != option_map_.end()) {
#if defined(CYGWIN)
char* str_end = nullptr;
value = std::strtod(itr->second.c_str(), &str_end);
if (str_end == itr->second.c_str()) {
exec_state =
LDBCommandExecuteResult::Failed(option + " has an invalid value.");
} else if (errno == ERANGE) {
exec_state = LDBCommandExecuteResult::Failed(
option + " has a value out-of-range.");
} else {
return true;
}
#else
try {
value = std::stod(itr->second);
return true;
} catch (const std::invalid_argument&) {
exec_state =
LDBCommandExecuteResult::Failed(option + " has an invalid value.");
} catch (const std::out_of_range&) {
exec_state = LDBCommandExecuteResult::Failed(
option + " has a value out-of-range.");
}
#endif
}
return false;
}
/**
* Parses the specific integer option and fills in the value.
* Returns true if the option is found.
* Returns false if the option is not found or if there is an error parsing the
* value. If there is an error, the specified exec_state is also
* updated.
*/
bool LDBCommand::ParseIntOption(
const std::map<std::string, std::string>& /*options*/,
const std::string& option, int& value,
LDBCommandExecuteResult& exec_state) {
auto itr = option_map_.find(option);
if (itr != option_map_.end()) {
#if defined(CYGWIN)
char* str_end = nullptr;
value = strtol(itr->second.c_str(), &str_end, 10);
if (str_end == itr->second.c_str()) {
exec_state =
LDBCommandExecuteResult::Failed(option + " has an invalid value.");
} else if (errno == ERANGE) {
exec_state = LDBCommandExecuteResult::Failed(
option + " has a value out-of-range.");
} else {
return true;
}
#else
try {
value = std::stoi(itr->second);
return true;
} catch (const std::invalid_argument&) {
rocksdb: Replace ASSERT* with EXPECT* in functions that does not return void value Summary: gtest does not use exceptions to fail a unit test by design, and `ASSERT*`s are implemented using `return`. As a consequence we cannot use `ASSERT*` in a function that does not return `void` value ([[ https://code.google.com/p/googletest/wiki/AdvancedGuide#Assertion_Placement | 1]]), and have to fix our existing code. This diff does this in a generic way, with no manual changes. In order to detect all existing `ASSERT*` that are used in functions that doesn't return void value, I change the code to generate compile errors for such cases. In `util/testharness.h` I defined `EXPECT*` assertions, the same way as `ASSERT*`, and redefined `ASSERT*` to return `void`. Then executed: ```lang=bash % USE_CLANG=1 make all -j55 -k 2> build.log % perl -naF: -e 'print "-- -number=".$F[1]." ".$F[0]."\n" if /: error:/' \ build.log | xargs -L 1 perl -spi -e 's/ASSERT/EXPECT/g if $. == $number' % make format ``` After that I reverted back change to `ASSERT*` in `util/testharness.h`. But preserved introduced `EXPECT*`, which is the same as `ASSERT*`. This will be deleted once switched to gtest. This diff is independent and contains manual changes only in `util/testharness.h`. Test Plan: Make sure all tests are passing. ```lang=bash % USE_CLANG=1 make check ``` Reviewers: igor, lgalanis, sdong, yufei.zhu, rven, meyering Reviewed By: meyering Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D33333
2015-03-17 03:52:32 +00:00
exec_state =
LDBCommandExecuteResult::Failed(option + " has an invalid value.");
} catch (const std::out_of_range&) {
rocksdb: Replace ASSERT* with EXPECT* in functions that does not return void value Summary: gtest does not use exceptions to fail a unit test by design, and `ASSERT*`s are implemented using `return`. As a consequence we cannot use `ASSERT*` in a function that does not return `void` value ([[ https://code.google.com/p/googletest/wiki/AdvancedGuide#Assertion_Placement | 1]]), and have to fix our existing code. This diff does this in a generic way, with no manual changes. In order to detect all existing `ASSERT*` that are used in functions that doesn't return void value, I change the code to generate compile errors for such cases. In `util/testharness.h` I defined `EXPECT*` assertions, the same way as `ASSERT*`, and redefined `ASSERT*` to return `void`. Then executed: ```lang=bash % USE_CLANG=1 make all -j55 -k 2> build.log % perl -naF: -e 'print "-- -number=".$F[1]." ".$F[0]."\n" if /: error:/' \ build.log | xargs -L 1 perl -spi -e 's/ASSERT/EXPECT/g if $. == $number' % make format ``` After that I reverted back change to `ASSERT*` in `util/testharness.h`. But preserved introduced `EXPECT*`, which is the same as `ASSERT*`. This will be deleted once switched to gtest. This diff is independent and contains manual changes only in `util/testharness.h`. Test Plan: Make sure all tests are passing. ```lang=bash % USE_CLANG=1 make check ``` Reviewers: igor, lgalanis, sdong, yufei.zhu, rven, meyering Reviewed By: meyering Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D33333
2015-03-17 03:52:32 +00:00
exec_state = LDBCommandExecuteResult::Failed(
option + " has a value out-of-range.");
}
#endif
}
return false;
}
/**
* Parses the specified option and fills in the value.
* Returns true if the option is found.
* Returns false otherwise.
*/
bool LDBCommand::ParseStringOption(
const std::map<std::string, std::string>& /*options*/,
const std::string& option, std::string* value) {
auto itr = option_map_.find(option);
if (itr != option_map_.end()) {
*value = itr->second;
return true;
}
return false;
}
/**
* Parses the specified compression type and fills in the value.
* Returns true if the compression type is found.
* Returns false otherwise.
*/
bool LDBCommand::ParseCompressionTypeOption(
const std::map<std::string, std::string>& /*options*/,
const std::string& option, CompressionType& value,
LDBCommandExecuteResult& exec_state) {
auto itr = option_map_.find(option);
if (itr != option_map_.end()) {
const std::string& comp = itr->second;
if (comp == "no") {
value = kNoCompression;
return true;
} else if (comp == "snappy") {
value = kSnappyCompression;
return true;
} else if (comp == "zlib") {
value = kZlibCompression;
return true;
} else if (comp == "bzip2") {
value = kBZip2Compression;
return true;
} else if (comp == "lz4") {
value = kLZ4Compression;
return true;
} else if (comp == "lz4hc") {
value = kLZ4HCCompression;
return true;
} else if (comp == "xpress") {
value = kXpressCompression;
return true;
} else if (comp == "zstd") {
value = kZSTD;
return true;
} else {
// Unknown compression.
exec_state = LDBCommandExecuteResult::Failed(
"Unknown compression algorithm: " + comp);
}
}
return false;
}
void LDBCommand::OverrideBaseOptions() {
options_.create_if_missing = false;
int db_write_buffer_size;
if (ParseIntOption(option_map_, ARG_DB_WRITE_BUFFER_SIZE,
db_write_buffer_size, exec_state_)) {
if (db_write_buffer_size >= 0) {
options_.db_write_buffer_size = db_write_buffer_size;
} else {
exec_state_ = LDBCommandExecuteResult::Failed(ARG_DB_WRITE_BUFFER_SIZE +
" must be >= 0.");
}
}
if (options_.db_paths.size() == 0) {
options_.db_paths.emplace_back(db_path_,
std::numeric_limits<uint64_t>::max());
}
OverrideBaseCFOptions(static_cast<ColumnFamilyOptions*>(&options_));
}
void LDBCommand::OverrideBaseCFOptions(ColumnFamilyOptions* cf_opts) {
BlockBasedTableOptions table_options;
bool use_table_options = false;
int bits;
if (ParseIntOption(option_map_, ARG_BLOOM_BITS, bits, exec_state_)) {
if (bits > 0) {
use_table_options = true;
table_options.filter_policy.reset(NewBloomFilterPolicy(bits));
} else {
rocksdb: Replace ASSERT* with EXPECT* in functions that does not return void value Summary: gtest does not use exceptions to fail a unit test by design, and `ASSERT*`s are implemented using `return`. As a consequence we cannot use `ASSERT*` in a function that does not return `void` value ([[ https://code.google.com/p/googletest/wiki/AdvancedGuide#Assertion_Placement | 1]]), and have to fix our existing code. This diff does this in a generic way, with no manual changes. In order to detect all existing `ASSERT*` that are used in functions that doesn't return void value, I change the code to generate compile errors for such cases. In `util/testharness.h` I defined `EXPECT*` assertions, the same way as `ASSERT*`, and redefined `ASSERT*` to return `void`. Then executed: ```lang=bash % USE_CLANG=1 make all -j55 -k 2> build.log % perl -naF: -e 'print "-- -number=".$F[1]." ".$F[0]."\n" if /: error:/' \ build.log | xargs -L 1 perl -spi -e 's/ASSERT/EXPECT/g if $. == $number' % make format ``` After that I reverted back change to `ASSERT*` in `util/testharness.h`. But preserved introduced `EXPECT*`, which is the same as `ASSERT*`. This will be deleted once switched to gtest. This diff is independent and contains manual changes only in `util/testharness.h`. Test Plan: Make sure all tests are passing. ```lang=bash % USE_CLANG=1 make check ``` Reviewers: igor, lgalanis, sdong, yufei.zhu, rven, meyering Reviewed By: meyering Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D33333
2015-03-17 03:52:32 +00:00
exec_state_ =
LDBCommandExecuteResult::Failed(ARG_BLOOM_BITS + " must be > 0.");
}
}
int block_size;
if (ParseIntOption(option_map_, ARG_BLOCK_SIZE, block_size, exec_state_)) {
if (block_size > 0) {
use_table_options = true;
table_options.block_size = block_size;
} else {
rocksdb: Replace ASSERT* with EXPECT* in functions that does not return void value Summary: gtest does not use exceptions to fail a unit test by design, and `ASSERT*`s are implemented using `return`. As a consequence we cannot use `ASSERT*` in a function that does not return `void` value ([[ https://code.google.com/p/googletest/wiki/AdvancedGuide#Assertion_Placement | 1]]), and have to fix our existing code. This diff does this in a generic way, with no manual changes. In order to detect all existing `ASSERT*` that are used in functions that doesn't return void value, I change the code to generate compile errors for such cases. In `util/testharness.h` I defined `EXPECT*` assertions, the same way as `ASSERT*`, and redefined `ASSERT*` to return `void`. Then executed: ```lang=bash % USE_CLANG=1 make all -j55 -k 2> build.log % perl -naF: -e 'print "-- -number=".$F[1]." ".$F[0]."\n" if /: error:/' \ build.log | xargs -L 1 perl -spi -e 's/ASSERT/EXPECT/g if $. == $number' % make format ``` After that I reverted back change to `ASSERT*` in `util/testharness.h`. But preserved introduced `EXPECT*`, which is the same as `ASSERT*`. This will be deleted once switched to gtest. This diff is independent and contains manual changes only in `util/testharness.h`. Test Plan: Make sure all tests are passing. ```lang=bash % USE_CLANG=1 make check ``` Reviewers: igor, lgalanis, sdong, yufei.zhu, rven, meyering Reviewed By: meyering Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D33333
2015-03-17 03:52:32 +00:00
exec_state_ =
LDBCommandExecuteResult::Failed(ARG_BLOCK_SIZE + " must be > 0.");
}
}
cf_opts->force_consistency_checks = force_consistency_checks_;
if (use_table_options) {
cf_opts->table_factory.reset(NewBlockBasedTableFactory(table_options));
}
cf_opts->enable_blob_files = enable_blob_files_;
int min_blob_size;
if (ParseIntOption(option_map_, ARG_MIN_BLOB_SIZE, min_blob_size,
exec_state_)) {
if (min_blob_size >= 0) {
cf_opts->min_blob_size = min_blob_size;
} else {
exec_state_ =
LDBCommandExecuteResult::Failed(ARG_MIN_BLOB_SIZE + " must be >= 0.");
}
}
int blob_file_size;
if (ParseIntOption(option_map_, ARG_BLOB_FILE_SIZE, blob_file_size,
exec_state_)) {
if (blob_file_size > 0) {
cf_opts->blob_file_size = blob_file_size;
} else {
rocksdb: Replace ASSERT* with EXPECT* in functions that does not return void value Summary: gtest does not use exceptions to fail a unit test by design, and `ASSERT*`s are implemented using `return`. As a consequence we cannot use `ASSERT*` in a function that does not return `void` value ([[ https://code.google.com/p/googletest/wiki/AdvancedGuide#Assertion_Placement | 1]]), and have to fix our existing code. This diff does this in a generic way, with no manual changes. In order to detect all existing `ASSERT*` that are used in functions that doesn't return void value, I change the code to generate compile errors for such cases. In `util/testharness.h` I defined `EXPECT*` assertions, the same way as `ASSERT*`, and redefined `ASSERT*` to return `void`. Then executed: ```lang=bash % USE_CLANG=1 make all -j55 -k 2> build.log % perl -naF: -e 'print "-- -number=".$F[1]." ".$F[0]."\n" if /: error:/' \ build.log | xargs -L 1 perl -spi -e 's/ASSERT/EXPECT/g if $. == $number' % make format ``` After that I reverted back change to `ASSERT*` in `util/testharness.h`. But preserved introduced `EXPECT*`, which is the same as `ASSERT*`. This will be deleted once switched to gtest. This diff is independent and contains manual changes only in `util/testharness.h`. Test Plan: Make sure all tests are passing. ```lang=bash % USE_CLANG=1 make check ``` Reviewers: igor, lgalanis, sdong, yufei.zhu, rven, meyering Reviewed By: meyering Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D33333
2015-03-17 03:52:32 +00:00
exec_state_ =
LDBCommandExecuteResult::Failed(ARG_BLOB_FILE_SIZE + " must be > 0.");
}
}
cf_opts->enable_blob_garbage_collection = enable_blob_garbage_collection_;
double blob_garbage_collection_age_cutoff;
if (ParseDoubleOption(option_map_, ARG_BLOB_GARBAGE_COLLECTION_AGE_CUTOFF,
blob_garbage_collection_age_cutoff, exec_state_)) {
if (blob_garbage_collection_age_cutoff >= 0 &&
blob_garbage_collection_age_cutoff <= 1) {
cf_opts->blob_garbage_collection_age_cutoff =
blob_garbage_collection_age_cutoff;
} else {
exec_state_ = LDBCommandExecuteResult::Failed(
ARG_BLOB_GARBAGE_COLLECTION_AGE_CUTOFF + " must be >= 0 and <= 1.");
}
}
double blob_garbage_collection_force_threshold;
if (ParseDoubleOption(option_map_,
ARG_BLOB_GARBAGE_COLLECTION_FORCE_THRESHOLD,
blob_garbage_collection_force_threshold, exec_state_)) {
if (blob_garbage_collection_force_threshold >= 0 &&
blob_garbage_collection_force_threshold <= 1) {
cf_opts->blob_garbage_collection_force_threshold =
blob_garbage_collection_force_threshold;
} else {
exec_state_ = LDBCommandExecuteResult::Failed(
ARG_BLOB_GARBAGE_COLLECTION_FORCE_THRESHOLD +
" must be >= 0 and <= 1.");
}
}
int blob_compaction_readahead_size;
if (ParseIntOption(option_map_, ARG_BLOB_COMPACTION_READAHEAD_SIZE,
blob_compaction_readahead_size, exec_state_)) {
if (blob_compaction_readahead_size > 0) {
cf_opts->blob_compaction_readahead_size = blob_compaction_readahead_size;
} else {
exec_state_ = LDBCommandExecuteResult::Failed(
ARG_BLOB_COMPACTION_READAHEAD_SIZE + " must be > 0.");
}
}
Make it possible to enable blob files starting from a certain LSM tree level (#10077) Summary: Currently, if blob files are enabled (i.e. `enable_blob_files` is true), large values are extracted both during flush/recovery (when SST files are written into level 0 of the LSM tree) and during compaction into any LSM tree level. For certain use cases that have a mix of short-lived and long-lived values, it might make sense to support extracting large values only during compactions whose output level is greater than or equal to a specified LSM tree level (e.g. compactions into L1/L2/... or above). This could reduce the space amplification caused by large values that are turned into garbage shortly after being written at the price of some write amplification incurred by long-lived values whose extraction to blob files is delayed. In order to achieve this, we would like to do the following: - Add a new configuration option `blob_file_starting_level` (default: 0) to `AdvancedColumnFamilyOptions` (and `MutableCFOptions` and extend the related logic) - Instantiate `BlobFileBuilder` in `BuildTable` (used during flush and recovery, where the LSM tree level is L0) and `CompactionJob` iff `enable_blob_files` is set and the LSM tree level is `>= blob_file_starting_level` - Add unit tests for the new functionality, and add the new option to our stress tests (`db_stress` and `db_crashtest.py` ) - Add the new option to our benchmarking tool `db_bench` and the BlobDB benchmark script `run_blob_bench.sh` - Add the new option to the `ldb` tool (see https://github.com/facebook/rocksdb/wiki/Administration-and-Data-Access-Tool) - Ideally extend the C and Java bindings with the new option - Update the BlobDB wiki to document the new option. Pull Request resolved: https://github.com/facebook/rocksdb/pull/10077 Reviewed By: ltamasi Differential Revision: D36884156 Pulled By: gangliao fbshipit-source-id: 942bab025f04633edca8564ed64791cb5e31627d
2022-06-03 03:04:33 +00:00
int blob_file_starting_level;
if (ParseIntOption(option_map_, ARG_BLOB_FILE_STARTING_LEVEL,
blob_file_starting_level, exec_state_)) {
if (blob_file_starting_level >= 0) {
cf_opts->blob_file_starting_level = blob_file_starting_level;
} else {
exec_state_ = LDBCommandExecuteResult::Failed(
ARG_BLOB_FILE_STARTING_LEVEL + " must be >= 0.");
}
}
int prepopulate_blob_cache;
if (ParseIntOption(option_map_, ARG_PREPOPULATE_BLOB_CACHE,
prepopulate_blob_cache, exec_state_)) {
switch (prepopulate_blob_cache) {
case 0:
cf_opts->prepopulate_blob_cache = PrepopulateBlobCache::kDisable;
break;
case 1:
cf_opts->prepopulate_blob_cache = PrepopulateBlobCache::kFlushOnly;
break;
default:
exec_state_ = LDBCommandExecuteResult::Failed(
ARG_PREPOPULATE_BLOB_CACHE +
" must be 0 (disable) or 1 (flush only).");
}
}
Make it possible to enable blob files starting from a certain LSM tree level (#10077) Summary: Currently, if blob files are enabled (i.e. `enable_blob_files` is true), large values are extracted both during flush/recovery (when SST files are written into level 0 of the LSM tree) and during compaction into any LSM tree level. For certain use cases that have a mix of short-lived and long-lived values, it might make sense to support extracting large values only during compactions whose output level is greater than or equal to a specified LSM tree level (e.g. compactions into L1/L2/... or above). This could reduce the space amplification caused by large values that are turned into garbage shortly after being written at the price of some write amplification incurred by long-lived values whose extraction to blob files is delayed. In order to achieve this, we would like to do the following: - Add a new configuration option `blob_file_starting_level` (default: 0) to `AdvancedColumnFamilyOptions` (and `MutableCFOptions` and extend the related logic) - Instantiate `BlobFileBuilder` in `BuildTable` (used during flush and recovery, where the LSM tree level is L0) and `CompactionJob` iff `enable_blob_files` is set and the LSM tree level is `>= blob_file_starting_level` - Add unit tests for the new functionality, and add the new option to our stress tests (`db_stress` and `db_crashtest.py` ) - Add the new option to our benchmarking tool `db_bench` and the BlobDB benchmark script `run_blob_bench.sh` - Add the new option to the `ldb` tool (see https://github.com/facebook/rocksdb/wiki/Administration-and-Data-Access-Tool) - Ideally extend the C and Java bindings with the new option - Update the BlobDB wiki to document the new option. Pull Request resolved: https://github.com/facebook/rocksdb/pull/10077 Reviewed By: ltamasi Differential Revision: D36884156 Pulled By: gangliao fbshipit-source-id: 942bab025f04633edca8564ed64791cb5e31627d
2022-06-03 03:04:33 +00:00
auto itr = option_map_.find(ARG_AUTO_COMPACTION);
if (itr != option_map_.end()) {
cf_opts->disable_auto_compactions = !StringToBool(itr->second);
}
CompressionType compression_type;
if (ParseCompressionTypeOption(option_map_, ARG_COMPRESSION_TYPE,
compression_type, exec_state_)) {
cf_opts->compression = compression_type;
}
CompressionType blob_compression_type;
if (ParseCompressionTypeOption(option_map_, ARG_BLOB_COMPRESSION_TYPE,
blob_compression_type, exec_state_)) {
cf_opts->blob_compression_type = blob_compression_type;
}
int compression_max_dict_bytes;
if (ParseIntOption(option_map_, ARG_COMPRESSION_MAX_DICT_BYTES,
compression_max_dict_bytes, exec_state_)) {
if (compression_max_dict_bytes >= 0) {
cf_opts->compression_opts.max_dict_bytes = compression_max_dict_bytes;
} else {
exec_state_ = LDBCommandExecuteResult::Failed(
ARG_COMPRESSION_MAX_DICT_BYTES + " must be >= 0.");
}
}
int write_buffer_size;
if (ParseIntOption(option_map_, ARG_WRITE_BUFFER_SIZE, write_buffer_size,
exec_state_)) {
if (write_buffer_size > 0) {
cf_opts->write_buffer_size = write_buffer_size;
} else {
exec_state_ = LDBCommandExecuteResult::Failed(ARG_WRITE_BUFFER_SIZE +
rocksdb: Replace ASSERT* with EXPECT* in functions that does not return void value Summary: gtest does not use exceptions to fail a unit test by design, and `ASSERT*`s are implemented using `return`. As a consequence we cannot use `ASSERT*` in a function that does not return `void` value ([[ https://code.google.com/p/googletest/wiki/AdvancedGuide#Assertion_Placement | 1]]), and have to fix our existing code. This diff does this in a generic way, with no manual changes. In order to detect all existing `ASSERT*` that are used in functions that doesn't return void value, I change the code to generate compile errors for such cases. In `util/testharness.h` I defined `EXPECT*` assertions, the same way as `ASSERT*`, and redefined `ASSERT*` to return `void`. Then executed: ```lang=bash % USE_CLANG=1 make all -j55 -k 2> build.log % perl -naF: -e 'print "-- -number=".$F[1]." ".$F[0]."\n" if /: error:/' \ build.log | xargs -L 1 perl -spi -e 's/ASSERT/EXPECT/g if $. == $number' % make format ``` After that I reverted back change to `ASSERT*` in `util/testharness.h`. But preserved introduced `EXPECT*`, which is the same as `ASSERT*`. This will be deleted once switched to gtest. This diff is independent and contains manual changes only in `util/testharness.h`. Test Plan: Make sure all tests are passing. ```lang=bash % USE_CLANG=1 make check ``` Reviewers: igor, lgalanis, sdong, yufei.zhu, rven, meyering Reviewed By: meyering Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D33333
2015-03-17 03:52:32 +00:00
" must be > 0.");
}
}
int file_size;
if (ParseIntOption(option_map_, ARG_FILE_SIZE, file_size, exec_state_)) {
if (file_size > 0) {
cf_opts->target_file_size_base = file_size;
} else {
rocksdb: Replace ASSERT* with EXPECT* in functions that does not return void value Summary: gtest does not use exceptions to fail a unit test by design, and `ASSERT*`s are implemented using `return`. As a consequence we cannot use `ASSERT*` in a function that does not return `void` value ([[ https://code.google.com/p/googletest/wiki/AdvancedGuide#Assertion_Placement | 1]]), and have to fix our existing code. This diff does this in a generic way, with no manual changes. In order to detect all existing `ASSERT*` that are used in functions that doesn't return void value, I change the code to generate compile errors for such cases. In `util/testharness.h` I defined `EXPECT*` assertions, the same way as `ASSERT*`, and redefined `ASSERT*` to return `void`. Then executed: ```lang=bash % USE_CLANG=1 make all -j55 -k 2> build.log % perl -naF: -e 'print "-- -number=".$F[1]." ".$F[0]."\n" if /: error:/' \ build.log | xargs -L 1 perl -spi -e 's/ASSERT/EXPECT/g if $. == $number' % make format ``` After that I reverted back change to `ASSERT*` in `util/testharness.h`. But preserved introduced `EXPECT*`, which is the same as `ASSERT*`. This will be deleted once switched to gtest. This diff is independent and contains manual changes only in `util/testharness.h`. Test Plan: Make sure all tests are passing. ```lang=bash % USE_CLANG=1 make check ``` Reviewers: igor, lgalanis, sdong, yufei.zhu, rven, meyering Reviewed By: meyering Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D33333
2015-03-17 03:52:32 +00:00
exec_state_ =
LDBCommandExecuteResult::Failed(ARG_FILE_SIZE + " must be > 0.");
}
}
int fix_prefix_len;
if (ParseIntOption(option_map_, ARG_FIX_PREFIX_LEN, fix_prefix_len,
exec_state_)) {
if (fix_prefix_len > 0) {
cf_opts->prefix_extractor.reset(
NewFixedPrefixTransform(static_cast<size_t>(fix_prefix_len)));
} else {
exec_state_ =
LDBCommandExecuteResult::Failed(ARG_FIX_PREFIX_LEN + " must be > 0.");
}
}
}
Introduce a new storage specific Env API (#5761) Summary: The current Env API encompasses both storage/file operations, as well as OS related operations. Most of the APIs return a Status, which does not have enough metadata about an error, such as whether its retry-able or not, scope (i.e fault domain) of the error etc., that may be required in order to properly handle a storage error. The file APIs also do not provide enough control over the IO SLA, such as timeout, prioritization, hinting about placement and redundancy etc. This PR separates out the file/storage APIs from Env into a new FileSystem class. The APIs are updated to return an IOStatus with metadata about the error, as well as to take an IOOptions structure as input in order to allow more control over the IO. The user can set both ```options.env``` and ```options.file_system``` to specify that RocksDB should use the former for OS related operations and the latter for storage operations. Internally, a ```CompositeEnvWrapper``` has been introduced that inherits from ```Env``` and redirects individual methods to either an ```Env``` implementation or the ```FileSystem``` as appropriate. When options are sanitized during ```DB::Open```, ```options.env``` is replaced with a newly allocated ```CompositeEnvWrapper``` instance if both env and file_system have been specified. This way, the rest of the RocksDB code can continue to function as before. This PR also ports PosixEnv to the new API by splitting it into two - PosixEnv and PosixFileSystem. PosixEnv is defined as a sub-class of CompositeEnvWrapper, and threading/time functions are overridden with Posix specific implementations in order to avoid an extra level of indirection. The ```CompositeEnvWrapper``` translates ```IOStatus``` return code to ```Status```, and sets the severity to ```kSoftError``` if the io_status is retryable. The error handling code in RocksDB can then recover the DB automatically. Pull Request resolved: https://github.com/facebook/rocksdb/pull/5761 Differential Revision: D18868376 Pulled By: anand1976 fbshipit-source-id: 39efe18a162ea746fabac6360ff529baba48486f
2019-12-13 22:47:08 +00:00
// First, initializes the options state using the OPTIONS file when enabled.
// Second, overrides the options according to the CLI arguments and the
// specific subcommand being run.
void LDBCommand::PrepareOptions() {
if (!create_if_missing_ && try_load_options_) {
config_options_.env = options_.env;
Status s = LoadLatestOptions(config_options_, db_path_, &options_,
&column_families_);
if (!s.ok() && !s.IsNotFound()) {
// Option file exists but load option file error.
std::string current_version = std::to_string(ROCKSDB_MAJOR) + "." +
std::to_string(ROCKSDB_MINOR) + "." +
std::to_string(ROCKSDB_PATCH);
std::string msg =
s.ToString() + "\nThis tool was built with version " +
current_version +
". If your db is in a different version, please try again "
"with option --" +
LDBCommand::ARG_IGNORE_UNKNOWN_OPTIONS + ".";
exec_state_ = LDBCommandExecuteResult::Failed(msg);
db_ = nullptr;
return;
}
if (!options_.wal_dir.empty()) {
if (options_.env->FileExists(options_.wal_dir).IsNotFound()) {
options_.wal_dir = db_path_;
fprintf(
stderr,
"wal_dir loaded from the option file doesn't exist. Ignore it.\n");
}
}
// If merge operator is not set, set a string append operator.
for (auto& cf_entry : column_families_) {
if (!cf_entry.options.merge_operator) {
cf_entry.options.merge_operator =
MergeOperators::CreateStringAppendOperator(':');
}
}
}
if (options_.env == Env::Default()) {
options_.env = config_options_.env;
}
OverrideBaseOptions();
if (exec_state_.IsFailed()) {
return;
}
if (column_families_.empty()) {
// Reads the MANIFEST to figure out what column families exist. In this
// case, the option overrides from the CLI argument/specific subcommand
// apply to all column families.
std::vector<std::string> cf_list;
Status st = DB::ListColumnFamilies(options_, db_path_, &cf_list);
// It is possible the DB doesn't exist yet, for "create if not
// existing" case. The failure is ignored here. We rely on DB::Open()
// to give us the correct error message for problem with opening
// existing DB.
if (st.ok() && cf_list.size() > 1) {
// Ignore single column family DB.
for (auto cf_name : cf_list) {
column_families_.emplace_back(cf_name, options_);
}
}
} else {
// We got column families from the OPTIONS file. In this case, the option
// overrides from the CLI argument/specific subcommand only apply to the
// column family specified by `--column_family_name`.
auto column_families_iter =
std::find_if(column_families_.begin(), column_families_.end(),
[this](const ColumnFamilyDescriptor& cf_desc) {
return cf_desc.name == column_family_name_;
});
if (column_families_iter == column_families_.end()) {
exec_state_ = LDBCommandExecuteResult::Failed(
"Non-existing column family " + column_family_name_);
return;
}
OverrideBaseCFOptions(&column_families_iter->options);
}
}
bool LDBCommand::ParseKeyValue(const std::string& line, std::string* key,
std::string* value, bool is_key_hex,
bool is_value_hex) {
size_t pos = line.find(DELIM);
if (pos != std::string::npos) {
*key = line.substr(0, pos);
*value = line.substr(pos + strlen(DELIM));
if (is_key_hex) {
*key = HexToString(*key);
}
if (is_value_hex) {
*value = HexToString(*value);
}
return true;
} else {
return false;
}
}
/**
* Make sure that ONLY the command-line options and flags expected by this
* command are specified on the command-line. Extraneous options are usually
* the result of user error.
* Returns true if all checks pass. Else returns false, and prints an
* appropriate error msg to stderr.
*/
bool LDBCommand::ValidateCmdLineOptions() {
for (auto itr = option_map_.begin(); itr != option_map_.end(); ++itr) {
if (std::find(valid_cmd_line_options_.begin(),
valid_cmd_line_options_.end(),
itr->first) == valid_cmd_line_options_.end()) {
fprintf(stderr, "Invalid command-line option %s\n", itr->first.c_str());
return false;
}
}
for (std::vector<std::string>::const_iterator itr = flags_.begin();
itr != flags_.end(); ++itr) {
if (std::find(valid_cmd_line_options_.begin(),
valid_cmd_line_options_.end(),
*itr) == valid_cmd_line_options_.end()) {
fprintf(stderr, "Invalid command-line flag %s\n", itr->c_str());
return false;
}
}
if (!NoDBOpen() && option_map_.find(ARG_DB) == option_map_.end() &&
option_map_.find(ARG_PATH) == option_map_.end()) {
fprintf(stderr, "Either %s or %s must be specified.\n", ARG_DB.c_str(),
ARG_PATH.c_str());
return false;
}
return true;
}
std::string LDBCommand::HexToString(const std::string& str) {
std::string result;
std::string::size_type len = str.length();
if (len < 2 || str[0] != '0' || str[1] != 'x') {
fprintf(stderr, "Invalid hex input %s. Must start with 0x\n", str.c_str());
throw "Invalid hex input";
}
if (!Slice(str.data() + 2, len - 2).DecodeHex(&result)) {
throw "Invalid hex input";
}
return result;
}
std::string LDBCommand::StringToHex(const std::string& str) {
std::string result("0x");
result.append(Slice(str).ToString(true));
return result;
}
std::string LDBCommand::PrintKeyValue(const std::string& key,
const std::string& value, bool is_key_hex,
bool is_value_hex) {
std::string result;
result.append(is_key_hex ? StringToHex(key) : key);
result.append(DELIM);
result.append(is_value_hex ? StringToHex(value) : value);
return result;
}
std::string LDBCommand::PrintKeyValue(const std::string& key,
const std::string& value, bool is_hex) {
return PrintKeyValue(key, value, is_hex, is_hex);
}
Minor refactor on LDB command for wide column support and release note (#11777) Summary: As mentioned in https://github.com/facebook/rocksdb/issues/11754 , refactor to clean up some nearly identical logic. This PR changes the existing debugging string format of Scan command as the following. ``` ❯ ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/ scan --hex ``` Before ``` 0x6669727374 : :0x68656C6C6F 0x617474725F6E616D6531:0x666F6F 0x617474725F6E616D6532:0x626172 0x7365636F6E64 : 0x617474725F6F6E65:0x74776F 0x617474725F7468726565:0x666F7572 0x7468697264 : 0x62617A ``` After ``` 0x6669727374 ==> :0x68656C6C6F 0x617474725F6E616D6531:0x666F6F 0x617474725F6E616D6532:0x626172 0x7365636F6E64 ==> 0x617474725F6F6E65:0x74776F 0x617474725F7468726565:0x666F7572 0x7468697264 ==> 0x62617A ``` Pull Request resolved: https://github.com/facebook/rocksdb/pull/11777 Test Plan: ``` ❯ ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/ dump first ==> :hello attr_name1:foo attr_name2:bar second ==> attr_one:two attr_three:four third ==> baz Keys in range: 3 ❯ ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/ scan first ==> :hello attr_name1:foo attr_name2:bar second ==> attr_one:two attr_three:four third ==> baz ❯ ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/ dump --hex 0x6669727374 ==> :0x68656C6C6F 0x617474725F6E616D6531:0x666F6F 0x617474725F6E616D6532:0x626172 0x7365636F6E64 ==> 0x617474725F6F6E65:0x74776F 0x617474725F7468726565:0x666F7572 0x7468697264 ==> 0x62617A Keys in range: 3 ❯ ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/ scan --hex 0x6669727374 ==> :0x68656C6C6F 0x617474725F6E616D6531:0x666F6F 0x617474725F6E616D6532:0x626172 0x7365636F6E64 ==> 0x617474725F6F6E65:0x74776F 0x617474725F7468726565:0x666F7572 0x7468697264 ==> 0x62617A ``` Reviewed By: jowlyzhang Differential Revision: D48876755 Pulled By: jaykorean fbshipit-source-id: b1c608a810fe038999ac528b690d398abf5f21d7
2023-08-31 23:17:03 +00:00
std::string LDBCommand::PrintKeyValueOrWideColumns(
const Slice& key, const Slice& value, const WideColumns& wide_columns,
bool is_key_hex, bool is_value_hex) {
if (wide_columns.empty() ||
WideColumnsHelper::HasDefaultColumnOnly(wide_columns)) {
Minor refactor on LDB command for wide column support and release note (#11777) Summary: As mentioned in https://github.com/facebook/rocksdb/issues/11754 , refactor to clean up some nearly identical logic. This PR changes the existing debugging string format of Scan command as the following. ``` ❯ ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/ scan --hex ``` Before ``` 0x6669727374 : :0x68656C6C6F 0x617474725F6E616D6531:0x666F6F 0x617474725F6E616D6532:0x626172 0x7365636F6E64 : 0x617474725F6F6E65:0x74776F 0x617474725F7468726565:0x666F7572 0x7468697264 : 0x62617A ``` After ``` 0x6669727374 ==> :0x68656C6C6F 0x617474725F6E616D6531:0x666F6F 0x617474725F6E616D6532:0x626172 0x7365636F6E64 ==> 0x617474725F6F6E65:0x74776F 0x617474725F7468726565:0x666F7572 0x7468697264 ==> 0x62617A ``` Pull Request resolved: https://github.com/facebook/rocksdb/pull/11777 Test Plan: ``` ❯ ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/ dump first ==> :hello attr_name1:foo attr_name2:bar second ==> attr_one:two attr_three:four third ==> baz Keys in range: 3 ❯ ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/ scan first ==> :hello attr_name1:foo attr_name2:bar second ==> attr_one:two attr_three:four third ==> baz ❯ ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/ dump --hex 0x6669727374 ==> :0x68656C6C6F 0x617474725F6E616D6531:0x666F6F 0x617474725F6E616D6532:0x626172 0x7365636F6E64 ==> 0x617474725F6F6E65:0x74776F 0x617474725F7468726565:0x666F7572 0x7468697264 ==> 0x62617A Keys in range: 3 ❯ ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/ scan --hex 0x6669727374 ==> :0x68656C6C6F 0x617474725F6E616D6531:0x666F6F 0x617474725F6E616D6532:0x626172 0x7365636F6E64 ==> 0x617474725F6F6E65:0x74776F 0x617474725F7468726565:0x666F7572 0x7468697264 ==> 0x62617A ``` Reviewed By: jowlyzhang Differential Revision: D48876755 Pulled By: jaykorean fbshipit-source-id: b1c608a810fe038999ac528b690d398abf5f21d7
2023-08-31 23:17:03 +00:00
return PrintKeyValue(key.ToString(), value.ToString(), is_key_hex,
is_value_hex);
}
/*
// Sample plaintext output (first column is kDefaultWideColumnName)
key_1 ==> :foo attr_name1:bar attr_name2:baz
// Sample hex output (first column is kDefaultWideColumnName)
0x6669727374 ==> :0x68656C6C6F 0x617474725F6E616D6531:0x666F6F
*/
std::ostringstream oss;
WideColumnsHelper::DumpWideColumns(wide_columns, oss, is_value_hex);
return PrintKeyValue(key.ToString(), oss.str().c_str(), is_key_hex,
false); // is_value_hex_ is already honored in oss.
// avoid double-hexing it.
}
std::string LDBCommand::HelpRangeCmdArgs() {
std::ostringstream str_stream;
str_stream << " ";
str_stream << "[--" << ARG_FROM << "] ";
str_stream << "[--" << ARG_TO << "] ";
return str_stream.str();
}
bool LDBCommand::IsKeyHex(const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags) {
return (IsFlagPresent(flags, ARG_HEX) || IsFlagPresent(flags, ARG_KEY_HEX) ||
ParseBooleanOption(options, ARG_HEX, false) ||
ParseBooleanOption(options, ARG_KEY_HEX, false));
}
bool LDBCommand::IsValueHex(const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags) {
return (IsFlagPresent(flags, ARG_HEX) ||
IsFlagPresent(flags, ARG_VALUE_HEX) ||
ParseBooleanOption(options, ARG_HEX, false) ||
ParseBooleanOption(options, ARG_VALUE_HEX, false));
}
bool LDBCommand::IsTryLoadOptions(
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags) {
if (IsFlagPresent(flags, ARG_TRY_LOAD_OPTIONS)) {
return true;
}
// if `DB` is specified and not explicitly to create a new db, default
// `try_load_options` to true. The user could still disable that by set
// `try_load_options=false`.
// Note: Opening as TTL DB doesn't support `try_load_options`, so it's default
// to false. TODO: TTL_DB may need to fix that, otherwise it's unable to open
// DB which has incompatible setting with default options.
bool default_val = (options.find(ARG_DB) != options.end()) &&
!IsFlagPresent(flags, ARG_CREATE_IF_MISSING) &&
!IsFlagPresent(flags, ARG_TTL);
return ParseBooleanOption(options, ARG_TRY_LOAD_OPTIONS, default_val);
}
bool LDBCommand::ParseBooleanOption(
const std::map<std::string, std::string>& options,
const std::string& option, bool default_val) {
auto itr = options.find(option);
if (itr != options.end()) {
std::string option_val = itr->second;
return StringToBool(itr->second);
}
return default_val;
}
bool LDBCommand::StringToBool(std::string val) {
std::transform(val.begin(), val.end(), val.begin(),
[](char ch) -> char { return (char)::tolower(ch); });
if (val == "true") {
return true;
} else if (val == "false") {
return false;
} else {
throw "Invalid value for boolean argument";
}
}
CompactorCommand::CompactorCommand(
const std::vector<std::string>& /*params*/,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(options, flags, false,
BuildCmdLineOptions({ARG_FROM, ARG_TO, ARG_HEX, ARG_KEY_HEX,
ARG_VALUE_HEX, ARG_TTL})),
null_from_(true),
null_to_(true) {
auto itr = options.find(ARG_FROM);
if (itr != options.end()) {
null_from_ = false;
from_ = itr->second;
}
itr = options.find(ARG_TO);
if (itr != options.end()) {
null_to_ = false;
to_ = itr->second;
}
if (is_key_hex_) {
if (!null_from_) {
from_ = HexToString(from_);
}
if (!null_to_) {
to_ = HexToString(to_);
}
}
}
void CompactorCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(CompactorCommand::Name());
ret.append(HelpRangeCmdArgs());
ret.append("\n");
}
void CompactorCommand::DoCommand() {
if (!db_) {
assert(GetExecuteState().IsFailed());
return;
}
Slice* begin = nullptr;
Slice* end = nullptr;
if (!null_from_) {
begin = new Slice(from_);
}
if (!null_to_) {
end = new Slice(to_);
}
CompactRangeOptions cro;
cro.bottommost_level_compaction = BottommostLevelCompaction::kForceOptimized;
Status s = db_->CompactRange(cro, GetCfHandle(), begin, end);
if (!s.ok()) {
std::stringstream oss;
oss << "Compaction failed: " << s.ToString();
exec_state_ = LDBCommandExecuteResult::Failed(oss.str());
} else {
exec_state_ = LDBCommandExecuteResult::Succeed("");
}
delete begin;
delete end;
}
// ---------------------------------------------------------------------------
const std::string DBLoaderCommand::ARG_DISABLE_WAL = "disable_wal";
const std::string DBLoaderCommand::ARG_BULK_LOAD = "bulk_load";
const std::string DBLoaderCommand::ARG_COMPACT = "compact";
DBLoaderCommand::DBLoaderCommand(
const std::vector<std::string>& /*params*/,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(
options, flags, false,
BuildCmdLineOptions({ARG_HEX, ARG_KEY_HEX, ARG_VALUE_HEX, ARG_FROM,
ARG_TO, ARG_CREATE_IF_MISSING, ARG_DISABLE_WAL,
ARG_BULK_LOAD, ARG_COMPACT})),
disable_wal_(false),
bulk_load_(false),
compact_(false) {
create_if_missing_ = IsFlagPresent(flags, ARG_CREATE_IF_MISSING);
disable_wal_ = IsFlagPresent(flags, ARG_DISABLE_WAL);
bulk_load_ = IsFlagPresent(flags, ARG_BULK_LOAD);
compact_ = IsFlagPresent(flags, ARG_COMPACT);
}
void DBLoaderCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(DBLoaderCommand::Name());
ret.append(" [--" + ARG_CREATE_IF_MISSING + "]");
ret.append(" [--" + ARG_DISABLE_WAL + "]");
ret.append(" [--" + ARG_BULK_LOAD + "]");
ret.append(" [--" + ARG_COMPACT + "]");
ret.append("\n");
}
void DBLoaderCommand::OverrideBaseOptions() {
LDBCommand::OverrideBaseOptions();
options_.create_if_missing = create_if_missing_;
if (bulk_load_) {
options_.PrepareForBulkLoad();
}
}
void DBLoaderCommand::DoCommand() {
if (!db_) {
assert(GetExecuteState().IsFailed());
return;
}
WriteOptions write_options;
if (disable_wal_) {
write_options.disableWAL = true;
}
int bad_lines = 0;
std::string line;
// prefer ifstream getline performance vs that from std::cin istream
std::ifstream ifs_stdin("/dev/stdin");
std::istream* istream_p = ifs_stdin.is_open() ? &ifs_stdin : &std::cin;
Status s;
while (s.ok() && getline(*istream_p, line, '\n')) {
std::string key;
std::string value;
if (ParseKeyValue(line, &key, &value, is_key_hex_, is_value_hex_)) {
s = db_->Put(write_options, GetCfHandle(), Slice(key), Slice(value));
} else if (0 == line.find("Keys in range:")) {
// ignore this line
} else if (0 == line.find("Created bg thread 0x")) {
// ignore this line
} else {
bad_lines++;
}
}
if (bad_lines > 0) {
std::cout << "Warning: " << bad_lines << " bad lines ignored." << std::endl;
}
if (!s.ok()) {
std::stringstream oss;
oss << "Load failed: " << s.ToString();
exec_state_ = LDBCommandExecuteResult::Failed(oss.str());
}
if (compact_ && s.ok()) {
s = db_->CompactRange(CompactRangeOptions(), GetCfHandle(), nullptr,
nullptr);
}
if (!s.ok()) {
std::stringstream oss;
oss << "Compaction failed: " << s.ToString();
exec_state_ = LDBCommandExecuteResult::Failed(oss.str());
}
}
// ----------------------------------------------------------------------------
namespace {
void DumpManifestFile(Options options, std::string file, bool verbose, bool hex,
bool json,
const std::vector<ColumnFamilyDescriptor>& cf_descs) {
EnvOptions sopt;
std::string dbname("dummy");
std::shared_ptr<Cache> tc(NewLRUCache(options.max_open_files - 10,
options.table_cache_numshardbits));
// Notice we are using the default options not through SanitizeOptions(),
// if VersionSet::DumpManifest() depends on any option done by
// SanitizeOptions(), we need to initialize it manually.
options.db_paths.emplace_back("dummy", 0);
options.num_levels = 64;
WriteController wc(options.delayed_write_rate);
WriteBufferManager wb(options.db_write_buffer_size);
ImmutableDBOptions immutable_db_options(options);
VersionSet versions(dbname, &immutable_db_options, sopt, tc.get(), &wb, &wc,
/*block_cache_tracer=*/nullptr, /*io_tracer=*/nullptr,
/*db_id*/ "", /*db_session_id*/ "");
Status s = versions.DumpManifest(options, file, verbose, hex, json, cf_descs);
if (!s.ok()) {
fprintf(stderr, "Error in processing file %s %s\n", file.c_str(),
s.ToString().c_str());
}
}
} // namespace
const std::string ManifestDumpCommand::ARG_VERBOSE = "verbose";
const std::string ManifestDumpCommand::ARG_JSON = "json";
const std::string ManifestDumpCommand::ARG_PATH = "path";
void ManifestDumpCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(ManifestDumpCommand::Name());
ret.append(" [--" + ARG_VERBOSE + "]");
Added JSON manifest dump option to ldb command Summary: Added a new flag --json to the ldb manifest_dump command that prints out the version edits as JSON objects for easier reading and parsing of information. Test Plan: **Sample usage: ** ``` ./ldb manifest_dump --json --path=path/to/manifest/file ``` **Sample output:** ``` {"EditNumber": 0, "Comparator": "leveldb.BytewiseComparator", "ColumnFamily": 0} {"EditNumber": 1, "LogNumber": 0, "ColumnFamily": 0} {"EditNumber": 2, "LogNumber": 4, "PrevLogNumber": 0, "NextFileNumber": 7, "LastSeq": 35356, "AddedFiles": [{"Level": 0, "FileNumber": 5, "FileSize": 1949284, "SmallestIKey": "'", "LargestIKey": "'"}], "ColumnFamily": 0} ... {"EditNumber": 13, "PrevLogNumber": 0, "NextFileNumber": 36, "LastSeq": 290994, "DeletedFiles": [{"Level": 0, "FileNumber": 17}, {"Level": 0, "FileNumber": 20}, {"Level": 0, "FileNumber": 22}, {"Level": 0, "FileNumber": 24}, {"Level": 1, "FileNumber": 13}, {"Level": 1, "FileNumber": 14}, {"Level": 1, "FileNumber": 15}, {"Level": 1, "FileNumber": 18}], "AddedFiles": [{"Level": 1, "FileNumber": 25, "FileSize": 2114340, "SmallestIKey": "'", "LargestIKey": "'"}, {"Level": 1, "FileNumber": 26, "FileSize": 2115213, "SmallestIKey": "'", "LargestIKey": "'"}, {"Level": 1, "FileNumber": 27, "FileSize": 2114807, "SmallestIKey": "'", "LargestIKey": "'"}, {"Level": 1, "FileNumber": 30, "FileSize": 2115271, "SmallestIKey": "'", "LargestIKey": "'"}, {"Level": 1, "FileNumber": 31, "FileSize": 2115165, "SmallestIKey": "'", "LargestIKey": "'"}, {"Level": 1, "FileNumber": 32, "FileSize": 2114683, "SmallestIKey": "'", "LargestIKey": "'"}, {"Level": 1, "FileNumber": 35, "FileSize": 1757512, "SmallestIKey": "'", "LargestIKey": "'"}], "ColumnFamily": 0} ... ``` Reviewers: sdong, anthony, yhchiang, igor Reviewed By: igor Subscribers: dhruba Differential Revision: https://reviews.facebook.net/D41727
2015-07-17 17:07:40 +00:00
ret.append(" [--" + ARG_JSON + "]");
ret.append(" [--" + ARG_PATH + "=<path_to_manifest_file>]");
ret.append("\n");
}
ManifestDumpCommand::ManifestDumpCommand(
const std::vector<std::string>& /*params*/,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(
options, flags, false,
BuildCmdLineOptions({ARG_VERBOSE, ARG_PATH, ARG_HEX, ARG_JSON})),
verbose_(false),
json_(false),
path_("") {
verbose_ = IsFlagPresent(flags, ARG_VERBOSE);
Added JSON manifest dump option to ldb command Summary: Added a new flag --json to the ldb manifest_dump command that prints out the version edits as JSON objects for easier reading and parsing of information. Test Plan: **Sample usage: ** ``` ./ldb manifest_dump --json --path=path/to/manifest/file ``` **Sample output:** ``` {"EditNumber": 0, "Comparator": "leveldb.BytewiseComparator", "ColumnFamily": 0} {"EditNumber": 1, "LogNumber": 0, "ColumnFamily": 0} {"EditNumber": 2, "LogNumber": 4, "PrevLogNumber": 0, "NextFileNumber": 7, "LastSeq": 35356, "AddedFiles": [{"Level": 0, "FileNumber": 5, "FileSize": 1949284, "SmallestIKey": "'", "LargestIKey": "'"}], "ColumnFamily": 0} ... {"EditNumber": 13, "PrevLogNumber": 0, "NextFileNumber": 36, "LastSeq": 290994, "DeletedFiles": [{"Level": 0, "FileNumber": 17}, {"Level": 0, "FileNumber": 20}, {"Level": 0, "FileNumber": 22}, {"Level": 0, "FileNumber": 24}, {"Level": 1, "FileNumber": 13}, {"Level": 1, "FileNumber": 14}, {"Level": 1, "FileNumber": 15}, {"Level": 1, "FileNumber": 18}], "AddedFiles": [{"Level": 1, "FileNumber": 25, "FileSize": 2114340, "SmallestIKey": "'", "LargestIKey": "'"}, {"Level": 1, "FileNumber": 26, "FileSize": 2115213, "SmallestIKey": "'", "LargestIKey": "'"}, {"Level": 1, "FileNumber": 27, "FileSize": 2114807, "SmallestIKey": "'", "LargestIKey": "'"}, {"Level": 1, "FileNumber": 30, "FileSize": 2115271, "SmallestIKey": "'", "LargestIKey": "'"}, {"Level": 1, "FileNumber": 31, "FileSize": 2115165, "SmallestIKey": "'", "LargestIKey": "'"}, {"Level": 1, "FileNumber": 32, "FileSize": 2114683, "SmallestIKey": "'", "LargestIKey": "'"}, {"Level": 1, "FileNumber": 35, "FileSize": 1757512, "SmallestIKey": "'", "LargestIKey": "'"}], "ColumnFamily": 0} ... ``` Reviewers: sdong, anthony, yhchiang, igor Reviewed By: igor Subscribers: dhruba Differential Revision: https://reviews.facebook.net/D41727
2015-07-17 17:07:40 +00:00
json_ = IsFlagPresent(flags, ARG_JSON);
auto itr = options.find(ARG_PATH);
if (itr != options.end()) {
path_ = itr->second;
if (path_.empty()) {
exec_state_ = LDBCommandExecuteResult::Failed("--path: missing pathname");
}
}
}
void ManifestDumpCommand::DoCommand() {
PrepareOptions();
std::string manifestfile;
if (!path_.empty()) {
manifestfile = path_;
} else {
// We need to find the manifest file by searching the directory
// containing the db for files of the form MANIFEST_[0-9]+
std::vector<std::string> files;
Status s = options_.env->GetChildren(db_path_, &files);
if (!s.ok()) {
std::string err_msg = s.ToString();
err_msg.append(": Failed to list the content of ");
err_msg.append(db_path_);
exec_state_ = LDBCommandExecuteResult::Failed(err_msg);
return;
}
const std::string kManifestNamePrefix = "MANIFEST-";
std::string matched_file;
#ifdef OS_WIN
const char kPathDelim = '\\';
#else
const char kPathDelim = '/';
#endif
for (const auto& file_path : files) {
// Some Env::GetChildren() return absolute paths. Some directories' path
// end with path delim, e.g. '/' or '\\'.
size_t pos = file_path.find_last_of(kPathDelim);
if (pos == file_path.size() - 1) {
continue;
}
std::string fname;
if (pos != std::string::npos) {
// Absolute path.
fname.assign(file_path, pos + 1, file_path.size() - pos - 1);
} else {
fname = file_path;
}
uint64_t file_num = 0;
FileType file_type = kWalFile; // Just for initialization
if (ParseFileName(fname, &file_num, &file_type) &&
file_type == kDescriptorFile) {
if (!matched_file.empty()) {
exec_state_ = LDBCommandExecuteResult::Failed(
rocksdb: Replace ASSERT* with EXPECT* in functions that does not return void value Summary: gtest does not use exceptions to fail a unit test by design, and `ASSERT*`s are implemented using `return`. As a consequence we cannot use `ASSERT*` in a function that does not return `void` value ([[ https://code.google.com/p/googletest/wiki/AdvancedGuide#Assertion_Placement | 1]]), and have to fix our existing code. This diff does this in a generic way, with no manual changes. In order to detect all existing `ASSERT*` that are used in functions that doesn't return void value, I change the code to generate compile errors for such cases. In `util/testharness.h` I defined `EXPECT*` assertions, the same way as `ASSERT*`, and redefined `ASSERT*` to return `void`. Then executed: ```lang=bash % USE_CLANG=1 make all -j55 -k 2> build.log % perl -naF: -e 'print "-- -number=".$F[1]." ".$F[0]."\n" if /: error:/' \ build.log | xargs -L 1 perl -spi -e 's/ASSERT/EXPECT/g if $. == $number' % make format ``` After that I reverted back change to `ASSERT*` in `util/testharness.h`. But preserved introduced `EXPECT*`, which is the same as `ASSERT*`. This will be deleted once switched to gtest. This diff is independent and contains manual changes only in `util/testharness.h`. Test Plan: Make sure all tests are passing. ```lang=bash % USE_CLANG=1 make check ``` Reviewers: igor, lgalanis, sdong, yufei.zhu, rven, meyering Reviewed By: meyering Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D33333
2015-03-17 03:52:32 +00:00
"Multiple MANIFEST files found; use --path to select one");
return;
} else {
matched_file.swap(fname);
}
}
}
if (matched_file.empty()) {
std::string err_msg("No MANIFEST found in ");
err_msg.append(db_path_);
exec_state_ = LDBCommandExecuteResult::Failed(err_msg);
return;
}
if (db_path_.back() != '/') {
db_path_.append("/");
}
manifestfile = db_path_ + matched_file;
}
if (verbose_) {
fprintf(stdout, "Processing Manifest file %s\n", manifestfile.c_str());
}
DumpManifestFile(options_, manifestfile, verbose_, is_key_hex_, json_,
column_families_);
Added JSON manifest dump option to ldb command Summary: Added a new flag --json to the ldb manifest_dump command that prints out the version edits as JSON objects for easier reading and parsing of information. Test Plan: **Sample usage: ** ``` ./ldb manifest_dump --json --path=path/to/manifest/file ``` **Sample output:** ``` {"EditNumber": 0, "Comparator": "leveldb.BytewiseComparator", "ColumnFamily": 0} {"EditNumber": 1, "LogNumber": 0, "ColumnFamily": 0} {"EditNumber": 2, "LogNumber": 4, "PrevLogNumber": 0, "NextFileNumber": 7, "LastSeq": 35356, "AddedFiles": [{"Level": 0, "FileNumber": 5, "FileSize": 1949284, "SmallestIKey": "'", "LargestIKey": "'"}], "ColumnFamily": 0} ... {"EditNumber": 13, "PrevLogNumber": 0, "NextFileNumber": 36, "LastSeq": 290994, "DeletedFiles": [{"Level": 0, "FileNumber": 17}, {"Level": 0, "FileNumber": 20}, {"Level": 0, "FileNumber": 22}, {"Level": 0, "FileNumber": 24}, {"Level": 1, "FileNumber": 13}, {"Level": 1, "FileNumber": 14}, {"Level": 1, "FileNumber": 15}, {"Level": 1, "FileNumber": 18}], "AddedFiles": [{"Level": 1, "FileNumber": 25, "FileSize": 2114340, "SmallestIKey": "'", "LargestIKey": "'"}, {"Level": 1, "FileNumber": 26, "FileSize": 2115213, "SmallestIKey": "'", "LargestIKey": "'"}, {"Level": 1, "FileNumber": 27, "FileSize": 2114807, "SmallestIKey": "'", "LargestIKey": "'"}, {"Level": 1, "FileNumber": 30, "FileSize": 2115271, "SmallestIKey": "'", "LargestIKey": "'"}, {"Level": 1, "FileNumber": 31, "FileSize": 2115165, "SmallestIKey": "'", "LargestIKey": "'"}, {"Level": 1, "FileNumber": 32, "FileSize": 2114683, "SmallestIKey": "'", "LargestIKey": "'"}, {"Level": 1, "FileNumber": 35, "FileSize": 1757512, "SmallestIKey": "'", "LargestIKey": "'"}], "ColumnFamily": 0} ... ``` Reviewers: sdong, anthony, yhchiang, igor Reviewed By: igor Subscribers: dhruba Differential Revision: https://reviews.facebook.net/D41727
2015-07-17 17:07:40 +00:00
if (verbose_) {
fprintf(stdout, "Processing Manifest file %s done\n", manifestfile.c_str());
}
}
// ----------------------------------------------------------------------------
namespace {
Status GetLiveFilesChecksumInfoFromVersionSet(Options options,
const std::string& db_path,
FileChecksumList* checksum_list) {
EnvOptions sopt;
Status s;
std::string dbname(db_path);
std::shared_ptr<Cache> tc(NewLRUCache(options.max_open_files - 10,
options.table_cache_numshardbits));
// Notice we are using the default options not through SanitizeOptions(),
// if VersionSet::GetLiveFilesChecksumInfo depends on any option done by
// SanitizeOptions(), we need to initialize it manually.
options.db_paths.emplace_back(db_path, 0);
options.num_levels = 64;
WriteController wc(options.delayed_write_rate);
WriteBufferManager wb(options.db_write_buffer_size);
ImmutableDBOptions immutable_db_options(options);
VersionSet versions(dbname, &immutable_db_options, sopt, tc.get(), &wb, &wc,
/*block_cache_tracer=*/nullptr, /*io_tracer=*/nullptr,
/*db_id*/ "", /*db_session_id*/ "");
std::vector<std::string> cf_name_list;
s = versions.ListColumnFamilies(&cf_name_list, db_path,
Simplify migration to FileSystem API (#6552) Summary: The current Env/FileSystem API separation has a couple of issues - 1. It requires the user to specify 2 options - ```Options::env``` and ```Options::file_system``` - which means they have to make code changes to benefit from the new APIs. Furthermore, there is a risk of accessing the same APIs in two different ways, through Env in the old way and through FileSystem in the new way. The two may not always match, for example, if env is ```PosixEnv``` and FileSystem is a custom implementation. Any stray RocksDB calls to env will use the ```PosixEnv``` implementation rather than the file_system implementation. 2. There needs to be a simple way for the FileSystem developer to instantiate an Env for backward compatibility purposes. This PR solves the above issues and simplifies the migration in the following ways - 1. Embed a shared_ptr to the ```FileSystem``` in the ```Env```, and remove ```Options::file_system``` as a configurable option. This way, no code changes will be required in application code to benefit from the new API. The default Env constructor uses a ```LegacyFileSystemWrapper``` as the embedded ```FileSystem```. 1a. - This also makes it more robust by ensuring that even if RocksDB has some stray calls to Env APIs rather than FileSystem, they will go through the same object and thus there is no risk of getting out of sync. 2. Provide a ```NewCompositeEnv()``` API that can be used to construct a PosixEnv with a custom FileSystem implementation. This eliminates an indirection to call Env APIs, and relieves the FileSystem developer of the burden of having to implement wrappers for the Env APIs. 3. Add a couple of missing FileSystem APIs - ```SanitizeEnvOptions()``` and ```NewLogger()``` Tests: 1. New unit tests 2. make check and make asan_check Pull Request resolved: https://github.com/facebook/rocksdb/pull/6552 Reviewed By: riversand963 Differential Revision: D20592038 Pulled By: anand1976 fbshipit-source-id: c3801ad4153f96d21d5a3ae26c92ba454d1bf1f7
2020-03-24 04:50:42 +00:00
immutable_db_options.fs.get());
if (s.ok()) {
std::vector<ColumnFamilyDescriptor> cf_list;
for (const auto& name : cf_name_list) {
cf_list.emplace_back(name, ColumnFamilyOptions(options));
}
s = versions.Recover(cf_list, true);
}
if (s.ok()) {
s = versions.GetLiveFilesChecksumInfo(checksum_list);
}
return s;
}
} // namespace
const std::string FileChecksumDumpCommand::ARG_PATH = "path";
void FileChecksumDumpCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(FileChecksumDumpCommand::Name());
ret.append(" [--" + ARG_PATH + "=<path_to_manifest_file>]");
ret.append("\n");
}
FileChecksumDumpCommand::FileChecksumDumpCommand(
const std::vector<std::string>& /*params*/,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(options, flags, false,
BuildCmdLineOptions({ARG_PATH, ARG_HEX})),
path_("") {
auto itr = options.find(ARG_PATH);
if (itr != options.end()) {
path_ = itr->second;
if (path_.empty()) {
exec_state_ = LDBCommandExecuteResult::Failed("--path: missing pathname");
}
}
is_checksum_hex_ = IsFlagPresent(flags, ARG_HEX);
}
void FileChecksumDumpCommand::DoCommand() {
PrepareOptions();
// print out the checksum information in the following format:
// sst file number, checksum function name, checksum value
// sst file number, checksum function name, checksum value
// ......
std::unique_ptr<FileChecksumList> checksum_list(NewFileChecksumList());
Status s = GetLiveFilesChecksumInfoFromVersionSet(options_, db_path_,
checksum_list.get());
if (s.ok() && checksum_list != nullptr) {
std::vector<uint64_t> file_numbers;
std::vector<std::string> checksums;
std::vector<std::string> checksum_func_names;
s = checksum_list->GetAllFileChecksums(&file_numbers, &checksums,
&checksum_func_names);
if (s.ok()) {
for (size_t i = 0; i < file_numbers.size(); i++) {
assert(i < file_numbers.size());
assert(i < checksums.size());
assert(i < checksum_func_names.size());
std::string checksum;
if (is_checksum_hex_) {
checksum = StringToHex(checksums[i]);
} else {
checksum = std::move(checksums[i]);
}
fprintf(stdout, "%" PRId64 ", %s, %s\n", file_numbers[i],
checksum_func_names[i].c_str(), checksum.c_str());
}
fprintf(stdout, "Print SST file checksum information finished \n");
}
}
if (!s.ok()) {
exec_state_ = LDBCommandExecuteResult::Failed(s.ToString());
}
}
// ----------------------------------------------------------------------------
aggregated-table-properties with GetMapProperty (#7779) Summary: So that we can more easily get aggregate live table data such as total filter, index, and data sizes. Also adds ldb support for getting properties Also fixed some missing/inaccurate related comments in db.h For example: $ ./ldb --db=testdb get_property rocksdb.aggregated-table-properties rocksdb.aggregated-table-properties.data_size: 102871 rocksdb.aggregated-table-properties.filter_size: 0 rocksdb.aggregated-table-properties.index_partitions: 0 rocksdb.aggregated-table-properties.index_size: 2232 rocksdb.aggregated-table-properties.num_data_blocks: 100 rocksdb.aggregated-table-properties.num_deletions: 0 rocksdb.aggregated-table-properties.num_entries: 15000 rocksdb.aggregated-table-properties.num_merge_operands: 0 rocksdb.aggregated-table-properties.num_range_deletions: 0 rocksdb.aggregated-table-properties.raw_key_size: 288890 rocksdb.aggregated-table-properties.raw_value_size: 198890 rocksdb.aggregated-table-properties.top_level_index_size: 0 $ ./ldb --db=testdb get_property rocksdb.aggregated-table-properties-at-level1 rocksdb.aggregated-table-properties-at-level1.data_size: 80909 rocksdb.aggregated-table-properties-at-level1.filter_size: 0 rocksdb.aggregated-table-properties-at-level1.index_partitions: 0 rocksdb.aggregated-table-properties-at-level1.index_size: 1787 rocksdb.aggregated-table-properties-at-level1.num_data_blocks: 81 rocksdb.aggregated-table-properties-at-level1.num_deletions: 0 rocksdb.aggregated-table-properties-at-level1.num_entries: 12466 rocksdb.aggregated-table-properties-at-level1.num_merge_operands: 0 rocksdb.aggregated-table-properties-at-level1.num_range_deletions: 0 rocksdb.aggregated-table-properties-at-level1.raw_key_size: 238210 rocksdb.aggregated-table-properties-at-level1.raw_value_size: 163414 rocksdb.aggregated-table-properties-at-level1.top_level_index_size: 0 $ Pull Request resolved: https://github.com/facebook/rocksdb/pull/7779 Test Plan: Added a test to ldb_test.py Reviewed By: jay-zhuang Differential Revision: D25653103 Pulled By: pdillinger fbshipit-source-id: 2905469a08a64dd6b5510cbd7be2e64d3234d6d3
2020-12-19 15:59:08 +00:00
void GetPropertyCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(GetPropertyCommand::Name());
ret.append(" <property_name>");
ret.append("\n");
}
GetPropertyCommand::GetPropertyCommand(
const std::vector<std::string>& params,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(options, flags, true, BuildCmdLineOptions({})) {
if (params.size() != 1) {
exec_state_ =
LDBCommandExecuteResult::Failed("property name must be specified");
} else {
property_ = params[0];
}
}
void GetPropertyCommand::DoCommand() {
if (!db_) {
assert(GetExecuteState().IsFailed());
return;
}
std::map<std::string, std::string> value_map;
std::string value;
// Rather than having different ldb command for map properties vs. string
// properties, we simply try Map property first. (This order only chosen
// because I prefer the map-style output for
// "rocksdb.aggregated-table-properties".)
if (db_->GetMapProperty(GetCfHandle(), property_, &value_map)) {
if (value_map.empty()) {
fprintf(stdout, "%s: <empty map>\n", property_.c_str());
} else {
for (auto& e : value_map) {
fprintf(stdout, "%s.%s: %s\n", property_.c_str(), e.first.c_str(),
e.second.c_str());
}
}
} else if (db_->GetProperty(GetCfHandle(), property_, &value)) {
fprintf(stdout, "%s: %s\n", property_.c_str(), value.c_str());
} else {
exec_state_ =
LDBCommandExecuteResult::Failed("failed to get property: " + property_);
}
}
// ----------------------------------------------------------------------------
void ListColumnFamiliesCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(ListColumnFamiliesCommand::Name());
ret.append("\n");
}
ListColumnFamiliesCommand::ListColumnFamiliesCommand(
const std::vector<std::string>& /*params*/,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(options, flags, false, BuildCmdLineOptions({})) {}
void ListColumnFamiliesCommand::DoCommand() {
PrepareOptions();
std::vector<std::string> column_families;
Status s = DB::ListColumnFamilies(options_, db_path_, &column_families);
if (!s.ok()) {
fprintf(stderr, "Error in processing db %s %s\n", db_path_.c_str(),
s.ToString().c_str());
} else {
fprintf(stdout, "Column families in %s: \n{", db_path_.c_str());
bool first = true;
for (auto cf : column_families) {
if (!first) {
fprintf(stdout, ", ");
}
first = false;
fprintf(stdout, "%s", cf.c_str());
}
fprintf(stdout, "}\n");
}
}
void CreateColumnFamilyCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(CreateColumnFamilyCommand::Name());
ret.append(" --db=<db_path> <new_column_family_name>");
ret.append("\n");
}
CreateColumnFamilyCommand::CreateColumnFamilyCommand(
const std::vector<std::string>& params,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(options, flags, true, {ARG_DB}) {
if (params.size() != 1) {
exec_state_ = LDBCommandExecuteResult::Failed(
"new column family name must be specified");
} else {
new_cf_name_ = params[0];
}
}
void CreateColumnFamilyCommand::DoCommand() {
if (!db_) {
assert(GetExecuteState().IsFailed());
return;
}
ColumnFamilyHandle* new_cf_handle = nullptr;
Status st = db_->CreateColumnFamily(options_, new_cf_name_, &new_cf_handle);
if (st.ok()) {
fprintf(stdout, "OK\n");
} else {
exec_state_ = LDBCommandExecuteResult::Failed(
"Fail to create new column family: " + st.ToString());
}
delete new_cf_handle;
CloseDB();
}
void DropColumnFamilyCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(DropColumnFamilyCommand::Name());
ret.append(" --db=<db_path> <column_family_name_to_drop>");
ret.append("\n");
}
DropColumnFamilyCommand::DropColumnFamilyCommand(
const std::vector<std::string>& params,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(options, flags, true, {ARG_DB}) {
if (params.size() != 1) {
exec_state_ = LDBCommandExecuteResult::Failed(
"The name of column family to drop must be specified");
} else {
cf_name_to_drop_ = params[0];
}
}
void DropColumnFamilyCommand::DoCommand() {
if (!db_) {
assert(GetExecuteState().IsFailed());
return;
}
auto iter = cf_handles_.find(cf_name_to_drop_);
if (iter == cf_handles_.end()) {
exec_state_ = LDBCommandExecuteResult::Failed(
"Column family: " + cf_name_to_drop_ + " doesn't exist in db.");
return;
}
ColumnFamilyHandle* cf_handle_to_drop = iter->second;
Status st = db_->DropColumnFamily(cf_handle_to_drop);
if (st.ok()) {
fprintf(stdout, "OK\n");
} else {
exec_state_ = LDBCommandExecuteResult::Failed(
"Fail to drop column family: " + st.ToString());
}
CloseDB();
}
// ----------------------------------------------------------------------------
namespace {
// This function only called when it's the sane case of >1 buckets in time-range
// Also called only when timekv falls between ttl_start and ttl_end provided
void IncBucketCounts(std::vector<uint64_t>& bucket_counts, int ttl_start,
int time_range, int bucket_size, int timekv,
int num_buckets) {
#ifdef NDEBUG
(void)time_range;
(void)num_buckets;
#endif
assert(time_range > 0 && timekv >= ttl_start && bucket_size > 0 &&
timekv < (ttl_start + time_range) && num_buckets > 1);
int bucket = (timekv - ttl_start) / bucket_size;
bucket_counts[bucket]++;
}
void PrintBucketCounts(const std::vector<uint64_t>& bucket_counts,
int ttl_start, int ttl_end, int bucket_size,
int num_buckets) {
int time_point = ttl_start;
for (int i = 0; i < num_buckets - 1; i++, time_point += bucket_size) {
fprintf(stdout, "Keys in range %s to %s : %lu\n",
TimeToHumanString(time_point).c_str(),
TimeToHumanString(time_point + bucket_size).c_str(),
(unsigned long)bucket_counts[i]);
}
fprintf(stdout, "Keys in range %s to %s : %lu\n",
TimeToHumanString(time_point).c_str(),
TimeToHumanString(ttl_end).c_str(),
(unsigned long)bucket_counts[num_buckets - 1]);
}
} // namespace
const std::string InternalDumpCommand::ARG_COUNT_ONLY = "count_only";
const std::string InternalDumpCommand::ARG_COUNT_DELIM = "count_delim";
const std::string InternalDumpCommand::ARG_STATS = "stats";
const std::string InternalDumpCommand::ARG_INPUT_KEY_HEX = "input_key_hex";
InternalDumpCommand::InternalDumpCommand(
const std::vector<std::string>& /*params*/,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(options, flags, true,
BuildCmdLineOptions(
{ARG_HEX, ARG_KEY_HEX, ARG_VALUE_HEX, ARG_FROM, ARG_TO,
ARG_MAX_KEYS, ARG_COUNT_ONLY, ARG_COUNT_DELIM, ARG_STATS,
ARG_INPUT_KEY_HEX, ARG_DECODE_BLOB_INDEX})),
has_from_(false),
has_to_(false),
max_keys_(-1),
delim_("."),
count_only_(false),
count_delim_(false),
print_stats_(false),
is_input_key_hex_(false),
decode_blob_index_(false) {
has_from_ = ParseStringOption(options, ARG_FROM, &from_);
has_to_ = ParseStringOption(options, ARG_TO, &to_);
ParseIntOption(options, ARG_MAX_KEYS, max_keys_, exec_state_);
auto itr = options.find(ARG_COUNT_DELIM);
if (itr != options.end()) {
delim_ = itr->second;
count_delim_ = true;
// fprintf(stdout,"delim = %c\n",delim_[0]);
} else {
count_delim_ = IsFlagPresent(flags, ARG_COUNT_DELIM);
delim_ = ".";
}
print_stats_ = IsFlagPresent(flags, ARG_STATS);
count_only_ = IsFlagPresent(flags, ARG_COUNT_ONLY);
is_input_key_hex_ = IsFlagPresent(flags, ARG_INPUT_KEY_HEX);
decode_blob_index_ = IsFlagPresent(flags, ARG_DECODE_BLOB_INDEX);
if (is_input_key_hex_) {
if (has_from_) {
from_ = HexToString(from_);
}
if (has_to_) {
to_ = HexToString(to_);
}
}
}
void InternalDumpCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(InternalDumpCommand::Name());
ret.append(HelpRangeCmdArgs());
ret.append(" [--" + ARG_INPUT_KEY_HEX + "]");
ret.append(" [--" + ARG_MAX_KEYS + "=<N>]");
ret.append(" [--" + ARG_COUNT_ONLY + "]");
ret.append(" [--" + ARG_COUNT_DELIM + "=<char>]");
ret.append(" [--" + ARG_STATS + "]");
ret.append(" [--" + ARG_DECODE_BLOB_INDEX + "]");
ret.append("\n");
}
void InternalDumpCommand::DoCommand() {
if (!db_) {
assert(GetExecuteState().IsFailed());
return;
}
if (print_stats_) {
std::string stats;
if (db_->GetProperty(GetCfHandle(), "rocksdb.stats", &stats)) {
fprintf(stdout, "%s\n", stats.c_str());
}
}
// Cast as DBImpl to get internal iterator
std::vector<KeyVersion> key_versions;
Status st = GetAllKeyVersions(db_, GetCfHandle(), from_, to_, max_keys_,
&key_versions);
if (!st.ok()) {
exec_state_ = LDBCommandExecuteResult::Failed(st.ToString());
return;
}
std::string rtype1, rtype2, row, val;
rtype2 = "";
uint64_t c = 0;
uint64_t s1 = 0, s2 = 0;
long long count = 0;
for (auto& key_version : key_versions) {
ValueType value_type = static_cast<ValueType>(key_version.type);
InternalKey ikey(key_version.user_key, key_version.sequence, value_type);
if (has_to_ && ikey.user_key() == to_) {
// GetAllKeyVersions() includes keys with user key `to_`, but idump has
// traditionally excluded such keys.
break;
}
++count;
int k;
if (count_delim_) {
rtype1 = "";
s1 = 0;
row = ikey.Encode().ToString();
val = key_version.value;
for (k = 0; row[k] != '\x01' && row[k] != '\0'; k++) s1++;
for (k = 0; val[k] != '\x01' && val[k] != '\0'; k++) s1++;
for (int j = 0; row[j] != delim_[0] && row[j] != '\0' && row[j] != '\x01';
j++)
rtype1 += row[j];
if (rtype2.compare("") && rtype2.compare(rtype1) != 0) {
fprintf(stdout, "%s => count:%" PRIu64 "\tsize:%" PRIu64 "\n",
rtype2.c_str(), c, s2);
c = 1;
s2 = s1;
rtype2 = rtype1;
} else {
c++;
s2 += s1;
rtype2 = rtype1;
}
}
if (!count_only_ && !count_delim_) {
std::string key = ikey.DebugString(is_key_hex_);
Slice value(key_version.value);
if (!decode_blob_index_ || value_type != kTypeBlobIndex) {
Wide Column support in ldb (#11754) Summary: wide_columns can now be pretty-printed in the following commands - `./ldb dump_wal` - `./ldb dump` - `./ldb idump` - `./ldb dump_live_files` - `./ldb scan` - `./sst_dump --command=scan` There are opportunities to refactor to reduce some nearly identical code. This PR is initial change to add wide column support in `ldb` and `sst_dump` tool. More PRs to come for the refactor. Pull Request resolved: https://github.com/facebook/rocksdb/pull/11754 Test Plan: **New Tests added** - `WideColumnsHelperTest::DumpWideColumns` - `WideColumnsHelperTest::DumpSliceAsWideColumns` **Changes added to existing tests** - `ExternalSSTFileTest::BasicMixed` added to cover mixed case (This test should have been added in https://github.com/facebook/rocksdb/issues/11688). This test does not verify the ldb or sst_dump output. This test was used to create test SST files having some rows with wide columns and some without and the generated SST files were used to manually test sst_dump_tool. - `createSST()` in `sst_dump_test` now takes `wide_column_one_in` to add wide column value in SST **dump_wal** ``` ./ldb dump_wal --walfile=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/000004.log --print_value --header ``` ``` Sequence,Count,ByteSize,Physical Offset,Key(s) : value 1,1,59,0,PUT_ENTITY(0) : 0x:0x68656C6C6F 0x617474725F6E616D6531:0x666F6F 0x617474725F6E616D6532:0x626172 2,1,34,42,PUT_ENTITY(0) : 0x617474725F6F6E65:0x74776F 0x617474725F7468726565:0x666F7572 3,1,17,7d,PUT(0) : 0x7468697264 : 0x62617A ``` **idump** ``` ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/ idump ``` ``` 'first' seq:1, type:22 => :hello attr_name1:foo attr_name2:bar 'second' seq:2, type:22 => attr_one:two attr_three:four 'third' seq:3, type:1 => baz Internal keys in range: 3 ``` **SST Dump from dump_live_files** ``` ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/ compact ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/ dump_live_files ``` ``` ... ============================== SST Files ============================== /tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/000013.sst level:1 ------------------------------ Process /tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/000013.sst Sst file format: block-based 'first' seq:0, type:22 => :hello attr_name1:foo attr_name2:bar 'second' seq:0, type:22 => attr_one:two attr_three:four 'third' seq:0, type:1 => baz ... ``` **dump** ``` ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/ dump ``` ``` first ==> :hello attr_name1:foo attr_name2:bar second ==> attr_one:two attr_three:four third ==> baz Keys in range: 3 ``` **scan** ``` ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/ scan ``` ``` first : :hello attr_name1:foo attr_name2:bar second : attr_one:two attr_three:four third : baz ``` **sst_dump** ``` ./sst_dump --file=/tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/000013.sst --command=scan ``` ``` options.env is 0x7ff54b296000 Process /tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/000013.sst Sst file format: block-based from [] to [] 'first' seq:0, type:22 => :hello attr_name1:foo attr_name2:bar 'second' seq:0, type:22 => attr_one:two attr_three:four 'third' seq:0, type:1 => baz ``` Reviewed By: ltamasi Differential Revision: D48837999 Pulled By: jaykorean fbshipit-source-id: b0280f0589d2b9716bb9b50530ffcabb397d140f
2023-08-30 19:45:52 +00:00
if (value_type == kTypeWideColumnEntity) {
std::ostringstream oss;
const Status s = WideColumnsHelper::DumpSliceAsWideColumns(
value, oss, is_value_hex_);
if (!s.ok()) {
fprintf(stderr, "%s => error deserializing wide columns\n",
key.c_str());
} else {
fprintf(stdout, "%s => %s\n", key.c_str(), oss.str().c_str());
}
} else {
fprintf(stdout, "%s => %s\n", key.c_str(),
value.ToString(is_value_hex_).c_str());
}
} else {
BlobIndex blob_index;
const Status s = blob_index.DecodeFrom(value);
if (!s.ok()) {
fprintf(stderr, "%s => error decoding blob index =>\n", key.c_str());
} else {
fprintf(stdout, "%s => %s\n", key.c_str(),
blob_index.DebugString(is_value_hex_).c_str());
}
}
}
// Terminate if maximum number of keys have been dumped
if (max_keys_ > 0 && count >= max_keys_) break;
}
if (count_delim_) {
fprintf(stdout, "%s => count:%" PRIu64 "\tsize:%" PRIu64 "\n",
rtype2.c_str(), c, s2);
} else {
fprintf(stdout, "Internal keys in range: %lld\n", count);
}
}
const std::string DBDumperCommand::ARG_COUNT_ONLY = "count_only";
const std::string DBDumperCommand::ARG_COUNT_DELIM = "count_delim";
const std::string DBDumperCommand::ARG_STATS = "stats";
const std::string DBDumperCommand::ARG_TTL_BUCKET = "bucket";
DBDumperCommand::DBDumperCommand(
const std::vector<std::string>& /*params*/,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(
options, flags, true,
BuildCmdLineOptions(
{ARG_TTL, ARG_HEX, ARG_KEY_HEX, ARG_VALUE_HEX, ARG_FROM, ARG_TO,
ARG_MAX_KEYS, ARG_COUNT_ONLY, ARG_COUNT_DELIM, ARG_STATS,
ARG_TTL_START, ARG_TTL_END, ARG_TTL_BUCKET, ARG_TIMESTAMP,
ARG_PATH, ARG_DECODE_BLOB_INDEX, ARG_DUMP_UNCOMPRESSED_BLOBS})),
null_from_(true),
null_to_(true),
max_keys_(-1),
count_only_(false),
count_delim_(false),
print_stats_(false),
decode_blob_index_(false) {
auto itr = options.find(ARG_FROM);
if (itr != options.end()) {
null_from_ = false;
from_ = itr->second;
}
itr = options.find(ARG_TO);
if (itr != options.end()) {
null_to_ = false;
to_ = itr->second;
}
itr = options.find(ARG_MAX_KEYS);
if (itr != options.end()) {
try {
#if defined(CYGWIN)
max_keys_ = strtol(itr->second.c_str(), 0, 10);
#else
max_keys_ = std::stoi(itr->second);
#endif
} catch (const std::invalid_argument&) {
exec_state_ = LDBCommandExecuteResult::Failed(ARG_MAX_KEYS +
rocksdb: Replace ASSERT* with EXPECT* in functions that does not return void value Summary: gtest does not use exceptions to fail a unit test by design, and `ASSERT*`s are implemented using `return`. As a consequence we cannot use `ASSERT*` in a function that does not return `void` value ([[ https://code.google.com/p/googletest/wiki/AdvancedGuide#Assertion_Placement | 1]]), and have to fix our existing code. This diff does this in a generic way, with no manual changes. In order to detect all existing `ASSERT*` that are used in functions that doesn't return void value, I change the code to generate compile errors for such cases. In `util/testharness.h` I defined `EXPECT*` assertions, the same way as `ASSERT*`, and redefined `ASSERT*` to return `void`. Then executed: ```lang=bash % USE_CLANG=1 make all -j55 -k 2> build.log % perl -naF: -e 'print "-- -number=".$F[1]." ".$F[0]."\n" if /: error:/' \ build.log | xargs -L 1 perl -spi -e 's/ASSERT/EXPECT/g if $. == $number' % make format ``` After that I reverted back change to `ASSERT*` in `util/testharness.h`. But preserved introduced `EXPECT*`, which is the same as `ASSERT*`. This will be deleted once switched to gtest. This diff is independent and contains manual changes only in `util/testharness.h`. Test Plan: Make sure all tests are passing. ```lang=bash % USE_CLANG=1 make check ``` Reviewers: igor, lgalanis, sdong, yufei.zhu, rven, meyering Reviewed By: meyering Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D33333
2015-03-17 03:52:32 +00:00
" has an invalid value");
} catch (const std::out_of_range&) {
rocksdb: Replace ASSERT* with EXPECT* in functions that does not return void value Summary: gtest does not use exceptions to fail a unit test by design, and `ASSERT*`s are implemented using `return`. As a consequence we cannot use `ASSERT*` in a function that does not return `void` value ([[ https://code.google.com/p/googletest/wiki/AdvancedGuide#Assertion_Placement | 1]]), and have to fix our existing code. This diff does this in a generic way, with no manual changes. In order to detect all existing `ASSERT*` that are used in functions that doesn't return void value, I change the code to generate compile errors for such cases. In `util/testharness.h` I defined `EXPECT*` assertions, the same way as `ASSERT*`, and redefined `ASSERT*` to return `void`. Then executed: ```lang=bash % USE_CLANG=1 make all -j55 -k 2> build.log % perl -naF: -e 'print "-- -number=".$F[1]." ".$F[0]."\n" if /: error:/' \ build.log | xargs -L 1 perl -spi -e 's/ASSERT/EXPECT/g if $. == $number' % make format ``` After that I reverted back change to `ASSERT*` in `util/testharness.h`. But preserved introduced `EXPECT*`, which is the same as `ASSERT*`. This will be deleted once switched to gtest. This diff is independent and contains manual changes only in `util/testharness.h`. Test Plan: Make sure all tests are passing. ```lang=bash % USE_CLANG=1 make check ``` Reviewers: igor, lgalanis, sdong, yufei.zhu, rven, meyering Reviewed By: meyering Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D33333
2015-03-17 03:52:32 +00:00
exec_state_ = LDBCommandExecuteResult::Failed(
ARG_MAX_KEYS + " has a value out-of-range");
}
}
itr = options.find(ARG_COUNT_DELIM);
if (itr != options.end()) {
delim_ = itr->second;
count_delim_ = true;
} else {
count_delim_ = IsFlagPresent(flags, ARG_COUNT_DELIM);
delim_ = ".";
}
print_stats_ = IsFlagPresent(flags, ARG_STATS);
count_only_ = IsFlagPresent(flags, ARG_COUNT_ONLY);
decode_blob_index_ = IsFlagPresent(flags, ARG_DECODE_BLOB_INDEX);
dump_uncompressed_blobs_ = IsFlagPresent(flags, ARG_DUMP_UNCOMPRESSED_BLOBS);
if (is_key_hex_) {
if (!null_from_) {
from_ = HexToString(from_);
}
if (!null_to_) {
to_ = HexToString(to_);
}
}
itr = options.find(ARG_PATH);
if (itr != options.end()) {
path_ = itr->second;
if (db_path_.empty()) {
db_path_ = path_;
}
}
}
void DBDumperCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(DBDumperCommand::Name());
ret.append(HelpRangeCmdArgs());
ret.append(" [--" + ARG_TTL + "]");
ret.append(" [--" + ARG_MAX_KEYS + "=<N>]");
ret.append(" [--" + ARG_TIMESTAMP + "]");
ret.append(" [--" + ARG_COUNT_ONLY + "]");
ret.append(" [--" + ARG_COUNT_DELIM + "=<char>]");
ret.append(" [--" + ARG_STATS + "]");
ret.append(" [--" + ARG_TTL_BUCKET + "=<N>]");
ret.append(" [--" + ARG_TTL_START + "=<N>:- is inclusive]");
ret.append(" [--" + ARG_TTL_END + "=<N>:- is exclusive]");
ret.append(" [--" + ARG_PATH + "=<path_to_a_file>]");
ret.append(" [--" + ARG_DECODE_BLOB_INDEX + "]");
ret.append(" [--" + ARG_DUMP_UNCOMPRESSED_BLOBS + "]");
ret.append("\n");
}
/**
* Handles two separate cases:
*
* 1) --db is specified - just dump the database.
*
* 2) --path is specified - determine based on file extension what dumping
* function to call. Please note that we intentionally use the extension
* and avoid probing the file contents under the assumption that renaming
* the files is not a supported scenario.
*
*/
void DBDumperCommand::DoCommand() {
if (!db_) {
assert(!path_.empty());
std::string fileName = GetFileNameFromPath(path_);
uint64_t number;
FileType type;
exec_state_ = LDBCommandExecuteResult::Succeed("");
if (!ParseFileName(fileName, &number, &type)) {
exec_state_ =
LDBCommandExecuteResult::Failed("Can't parse file type: " + path_);
return;
}
switch (type) {
case kWalFile:
// TODO(myabandeh): allow configuring is_write_commited
DumpWalFile(options_, path_, /* print_header_ */ true,
/* print_values_ */ true, true /* is_write_commited */,
&exec_state_);
break;
case kTableFile:
DumpSstFile(options_, path_, is_key_hex_, /* show_properties */ true,
decode_blob_index_, from_, to_);
break;
case kDescriptorFile:
DumpManifestFile(options_, path_, /* verbose_ */ false, is_key_hex_,
/* json_ */ false, column_families_);
break;
case kBlobFile:
DumpBlobFile(path_, is_key_hex_, is_value_hex_,
dump_uncompressed_blobs_);
break;
default:
exec_state_ = LDBCommandExecuteResult::Failed(
"File type not supported: " + path_);
break;
}
} else {
DoDumpCommand();
}
}
void DBDumperCommand::DoDumpCommand() {
assert(nullptr != db_);
assert(path_.empty());
// Parse command line args
uint64_t count = 0;
if (print_stats_) {
std::string stats;
if (db_->GetProperty("rocksdb.stats", &stats)) {
fprintf(stdout, "%s\n", stats.c_str());
}
}
// Setup key iterator
ReadOptions scan_read_opts;
scan_read_opts.total_order_seek = true;
Iterator* iter = db_->NewIterator(scan_read_opts, GetCfHandle());
Status st = iter->status();
if (!st.ok()) {
rocksdb: Replace ASSERT* with EXPECT* in functions that does not return void value Summary: gtest does not use exceptions to fail a unit test by design, and `ASSERT*`s are implemented using `return`. As a consequence we cannot use `ASSERT*` in a function that does not return `void` value ([[ https://code.google.com/p/googletest/wiki/AdvancedGuide#Assertion_Placement | 1]]), and have to fix our existing code. This diff does this in a generic way, with no manual changes. In order to detect all existing `ASSERT*` that are used in functions that doesn't return void value, I change the code to generate compile errors for such cases. In `util/testharness.h` I defined `EXPECT*` assertions, the same way as `ASSERT*`, and redefined `ASSERT*` to return `void`. Then executed: ```lang=bash % USE_CLANG=1 make all -j55 -k 2> build.log % perl -naF: -e 'print "-- -number=".$F[1]." ".$F[0]."\n" if /: error:/' \ build.log | xargs -L 1 perl -spi -e 's/ASSERT/EXPECT/g if $. == $number' % make format ``` After that I reverted back change to `ASSERT*` in `util/testharness.h`. But preserved introduced `EXPECT*`, which is the same as `ASSERT*`. This will be deleted once switched to gtest. This diff is independent and contains manual changes only in `util/testharness.h`. Test Plan: Make sure all tests are passing. ```lang=bash % USE_CLANG=1 make check ``` Reviewers: igor, lgalanis, sdong, yufei.zhu, rven, meyering Reviewed By: meyering Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D33333
2015-03-17 03:52:32 +00:00
exec_state_ =
LDBCommandExecuteResult::Failed("Iterator error." + st.ToString());
}
if (!null_from_) {
iter->Seek(from_);
} else {
iter->SeekToFirst();
}
int max_keys = max_keys_;
int ttl_start;
if (!ParseIntOption(option_map_, ARG_TTL_START, ttl_start, exec_state_)) {
ttl_start = DBWithTTLImpl::kMinTimestamp; // TTL introduction time
}
int ttl_end;
if (!ParseIntOption(option_map_, ARG_TTL_END, ttl_end, exec_state_)) {
ttl_end = DBWithTTLImpl::kMaxTimestamp; // Max time allowed by TTL feature
}
if (ttl_end < ttl_start) {
fprintf(stderr, "Error: End time can't be less than start time\n");
delete iter;
return;
}
int time_range = ttl_end - ttl_start;
int bucket_size;
if (!ParseIntOption(option_map_, ARG_TTL_BUCKET, bucket_size, exec_state_) ||
bucket_size <= 0) {
bucket_size = time_range; // Will have just 1 bucket by default
}
// cretaing variables for row count of each type
std::string rtype1, rtype2, row, val;
rtype2 = "";
uint64_t c = 0;
uint64_t s1 = 0, s2 = 0;
// At this point, bucket_size=0 => time_range=0
int num_buckets = (bucket_size >= time_range)
? 1
: ((time_range + bucket_size - 1) / bucket_size);
std::vector<uint64_t> bucket_counts(num_buckets, 0);
if (is_db_ttl_ && !count_only_ && timestamp_ && !count_delim_) {
fprintf(stdout, "Dumping key-values from %s to %s\n",
TimeToHumanString(ttl_start).c_str(),
TimeToHumanString(ttl_end).c_str());
}
HistogramImpl vsize_hist;
for (; iter->Valid(); iter->Next()) {
int rawtime = 0;
// If end marker was specified, we stop before it
if (!null_to_ && (iter->key().ToString() >= to_)) break;
// Terminate if maximum number of keys have been dumped
if (max_keys == 0) break;
if (is_db_ttl_) {
TtlIterator* it_ttl = static_cast_with_check<TtlIterator>(iter);
Iterator with timestamp (#6255) Summary: Preliminary support for iterator with user timestamp. Current implementation does not consider merge operator and reverse iterator. Auto compaction is also disabled in unit tests. Create an iterator with timestamp. ``` ... read_opts.timestamp = &ts; auto* iter = db->NewIterator(read_opts); // target is key without timestamp. for (iter->Seek(target); iter->Valid(); iter->Next()) {} for (iter->SeekToFirst(); iter->Valid(); iter->Next()) {} delete iter; read_opts.timestamp = &ts1; // lower_bound and upper_bound are without timestamp. read_opts.iterate_lower_bound = &lower_bound; read_opts.iterate_upper_bound = &upper_bound; auto* iter1 = db->NewIterator(read_opts); // Do Seek or SeekToFirst() delete iter1; ``` Test plan (dev server) ``` $make check ``` Simple benchmarking (dev server) 1. The overhead introduced by this PR even when timestamp is disabled. key size: 16 bytes value size: 100 bytes Entries: 1000000 Data reside in main memory, and try to stress iterator. Repeated three times on master and this PR. - Seek without next ``` ./db_bench -db=/dev/shm/rocksdbtest-1000 -benchmarks=fillseq,seekrandom -enable_pipelined_write=false -disable_wal=true -format_version=3 ``` master: 159047.0 ops/sec this PR: 158922.3 ops/sec (2% drop in throughput) - Seek and next 10 times ``` ./db_bench -db=/dev/shm/rocksdbtest-1000 -benchmarks=fillseq,seekrandom -enable_pipelined_write=false -disable_wal=true -format_version=3 -seek_nexts=10 ``` master: 109539.3 ops/sec this PR: 107519.7 ops/sec (2% drop in throughput) Pull Request resolved: https://github.com/facebook/rocksdb/pull/6255 Differential Revision: D19438227 Pulled By: riversand963 fbshipit-source-id: b66b4979486f8474619f4aa6bdd88598870b0746
2020-03-07 00:21:03 +00:00
rawtime = it_ttl->ttl_timestamp();
if (rawtime < ttl_start || rawtime >= ttl_end) {
continue;
}
}
if (max_keys > 0) {
--max_keys;
}
if (is_db_ttl_ && num_buckets > 1) {
IncBucketCounts(bucket_counts, ttl_start, time_range, bucket_size,
rawtime, num_buckets);
}
++count;
if (count_delim_) {
rtype1 = "";
row = iter->key().ToString();
val = iter->value().ToString();
s1 = row.size() + val.size();
for (int j = 0; row[j] != delim_[0] && row[j] != '\0'; j++)
rtype1 += row[j];
if (rtype2.compare("") && rtype2.compare(rtype1) != 0) {
fprintf(stdout, "%s => count:%" PRIu64 "\tsize:%" PRIu64 "\n",
rtype2.c_str(), c, s2);
c = 1;
s2 = s1;
rtype2 = rtype1;
} else {
c++;
s2 += s1;
rtype2 = rtype1;
}
}
if (count_only_) {
vsize_hist.Add(iter->value().size());
}
if (!count_only_ && !count_delim_) {
if (is_db_ttl_ && timestamp_) {
fprintf(stdout, "%s ", TimeToHumanString(rawtime).c_str());
}
Wide Column support in ldb (#11754) Summary: wide_columns can now be pretty-printed in the following commands - `./ldb dump_wal` - `./ldb dump` - `./ldb idump` - `./ldb dump_live_files` - `./ldb scan` - `./sst_dump --command=scan` There are opportunities to refactor to reduce some nearly identical code. This PR is initial change to add wide column support in `ldb` and `sst_dump` tool. More PRs to come for the refactor. Pull Request resolved: https://github.com/facebook/rocksdb/pull/11754 Test Plan: **New Tests added** - `WideColumnsHelperTest::DumpWideColumns` - `WideColumnsHelperTest::DumpSliceAsWideColumns` **Changes added to existing tests** - `ExternalSSTFileTest::BasicMixed` added to cover mixed case (This test should have been added in https://github.com/facebook/rocksdb/issues/11688). This test does not verify the ldb or sst_dump output. This test was used to create test SST files having some rows with wide columns and some without and the generated SST files were used to manually test sst_dump_tool. - `createSST()` in `sst_dump_test` now takes `wide_column_one_in` to add wide column value in SST **dump_wal** ``` ./ldb dump_wal --walfile=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/000004.log --print_value --header ``` ``` Sequence,Count,ByteSize,Physical Offset,Key(s) : value 1,1,59,0,PUT_ENTITY(0) : 0x:0x68656C6C6F 0x617474725F6E616D6531:0x666F6F 0x617474725F6E616D6532:0x626172 2,1,34,42,PUT_ENTITY(0) : 0x617474725F6F6E65:0x74776F 0x617474725F7468726565:0x666F7572 3,1,17,7d,PUT(0) : 0x7468697264 : 0x62617A ``` **idump** ``` ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/ idump ``` ``` 'first' seq:1, type:22 => :hello attr_name1:foo attr_name2:bar 'second' seq:2, type:22 => attr_one:two attr_three:four 'third' seq:3, type:1 => baz Internal keys in range: 3 ``` **SST Dump from dump_live_files** ``` ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/ compact ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/ dump_live_files ``` ``` ... ============================== SST Files ============================== /tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/000013.sst level:1 ------------------------------ Process /tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/000013.sst Sst file format: block-based 'first' seq:0, type:22 => :hello attr_name1:foo attr_name2:bar 'second' seq:0, type:22 => attr_one:two attr_three:four 'third' seq:0, type:1 => baz ... ``` **dump** ``` ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/ dump ``` ``` first ==> :hello attr_name1:foo attr_name2:bar second ==> attr_one:two attr_three:four third ==> baz Keys in range: 3 ``` **scan** ``` ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/ scan ``` ``` first : :hello attr_name1:foo attr_name2:bar second : attr_one:two attr_three:four third : baz ``` **sst_dump** ``` ./sst_dump --file=/tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/000013.sst --command=scan ``` ``` options.env is 0x7ff54b296000 Process /tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/000013.sst Sst file format: block-based from [] to [] 'first' seq:0, type:22 => :hello attr_name1:foo attr_name2:bar 'second' seq:0, type:22 => attr_one:two attr_three:four 'third' seq:0, type:1 => baz ``` Reviewed By: ltamasi Differential Revision: D48837999 Pulled By: jaykorean fbshipit-source-id: b0280f0589d2b9716bb9b50530ffcabb397d140f
2023-08-30 19:45:52 +00:00
// (TODO) TTL Iterator does not support wide columns yet.
Minor refactor on LDB command for wide column support and release note (#11777) Summary: As mentioned in https://github.com/facebook/rocksdb/issues/11754 , refactor to clean up some nearly identical logic. This PR changes the existing debugging string format of Scan command as the following. ``` ❯ ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/ scan --hex ``` Before ``` 0x6669727374 : :0x68656C6C6F 0x617474725F6E616D6531:0x666F6F 0x617474725F6E616D6532:0x626172 0x7365636F6E64 : 0x617474725F6F6E65:0x74776F 0x617474725F7468726565:0x666F7572 0x7468697264 : 0x62617A ``` After ``` 0x6669727374 ==> :0x68656C6C6F 0x617474725F6E616D6531:0x666F6F 0x617474725F6E616D6532:0x626172 0x7365636F6E64 ==> 0x617474725F6F6E65:0x74776F 0x617474725F7468726565:0x666F7572 0x7468697264 ==> 0x62617A ``` Pull Request resolved: https://github.com/facebook/rocksdb/pull/11777 Test Plan: ``` ❯ ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/ dump first ==> :hello attr_name1:foo attr_name2:bar second ==> attr_one:two attr_three:four third ==> baz Keys in range: 3 ❯ ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/ scan first ==> :hello attr_name1:foo attr_name2:bar second ==> attr_one:two attr_three:four third ==> baz ❯ ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/ dump --hex 0x6669727374 ==> :0x68656C6C6F 0x617474725F6E616D6531:0x666F6F 0x617474725F6E616D6532:0x626172 0x7365636F6E64 ==> 0x617474725F6F6E65:0x74776F 0x617474725F7468726565:0x666F7572 0x7468697264 ==> 0x62617A Keys in range: 3 ❯ ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/ scan --hex 0x6669727374 ==> :0x68656C6C6F 0x617474725F6E616D6531:0x666F6F 0x617474725F6E616D6532:0x626172 0x7365636F6E64 ==> 0x617474725F6F6E65:0x74776F 0x617474725F7468726565:0x666F7572 0x7468697264 ==> 0x62617A ``` Reviewed By: jowlyzhang Differential Revision: D48876755 Pulled By: jaykorean fbshipit-source-id: b1c608a810fe038999ac528b690d398abf5f21d7
2023-08-31 23:17:03 +00:00
std::string str =
is_db_ttl_
? PrintKeyValue(iter->key().ToString(), iter->value().ToString(),
is_key_hex_, is_value_hex_)
: PrintKeyValueOrWideColumns(iter->key(), iter->value(),
iter->columns(), is_key_hex_,
is_value_hex_);
fprintf(stdout, "%s\n", str.c_str());
}
}
if (num_buckets > 1 && is_db_ttl_) {
PrintBucketCounts(bucket_counts, ttl_start, ttl_end, bucket_size,
num_buckets);
} else if (count_delim_) {
fprintf(stdout, "%s => count:%" PRIu64 "\tsize:%" PRIu64 "\n",
rtype2.c_str(), c, s2);
} else {
fprintf(stdout, "Keys in range: %" PRIu64 "\n", count);
}
if (count_only_) {
fprintf(stdout, "Value size distribution: \n");
fprintf(stdout, "%s\n", vsize_hist.ToString().c_str());
}
// Clean up
delete iter;
}
const std::string ReduceDBLevelsCommand::ARG_NEW_LEVELS = "new_levels";
const std::string ReduceDBLevelsCommand::ARG_PRINT_OLD_LEVELS =
"print_old_levels";
ReduceDBLevelsCommand::ReduceDBLevelsCommand(
const std::vector<std::string>& /*params*/,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(options, flags, false,
BuildCmdLineOptions({ARG_NEW_LEVELS, ARG_PRINT_OLD_LEVELS})),
old_levels_(1 << 7),
new_levels_(-1),
print_old_levels_(false) {
ParseIntOption(option_map_, ARG_NEW_LEVELS, new_levels_, exec_state_);
print_old_levels_ = IsFlagPresent(flags, ARG_PRINT_OLD_LEVELS);
if (new_levels_ <= 0) {
exec_state_ = LDBCommandExecuteResult::Failed(
rocksdb: Replace ASSERT* with EXPECT* in functions that does not return void value Summary: gtest does not use exceptions to fail a unit test by design, and `ASSERT*`s are implemented using `return`. As a consequence we cannot use `ASSERT*` in a function that does not return `void` value ([[ https://code.google.com/p/googletest/wiki/AdvancedGuide#Assertion_Placement | 1]]), and have to fix our existing code. This diff does this in a generic way, with no manual changes. In order to detect all existing `ASSERT*` that are used in functions that doesn't return void value, I change the code to generate compile errors for such cases. In `util/testharness.h` I defined `EXPECT*` assertions, the same way as `ASSERT*`, and redefined `ASSERT*` to return `void`. Then executed: ```lang=bash % USE_CLANG=1 make all -j55 -k 2> build.log % perl -naF: -e 'print "-- -number=".$F[1]." ".$F[0]."\n" if /: error:/' \ build.log | xargs -L 1 perl -spi -e 's/ASSERT/EXPECT/g if $. == $number' % make format ``` After that I reverted back change to `ASSERT*` in `util/testharness.h`. But preserved introduced `EXPECT*`, which is the same as `ASSERT*`. This will be deleted once switched to gtest. This diff is independent and contains manual changes only in `util/testharness.h`. Test Plan: Make sure all tests are passing. ```lang=bash % USE_CLANG=1 make check ``` Reviewers: igor, lgalanis, sdong, yufei.zhu, rven, meyering Reviewed By: meyering Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D33333
2015-03-17 03:52:32 +00:00
" Use --" + ARG_NEW_LEVELS + " to specify a new level number\n");
}
}
std::vector<std::string> ReduceDBLevelsCommand::PrepareArgs(
const std::string& db_path, int new_levels, bool print_old_level) {
std::vector<std::string> ret;
ret.push_back("reduce_levels");
ret.push_back("--" + ARG_DB + "=" + db_path);
ret.push_back("--" + ARG_NEW_LEVELS + "=" + std::to_string(new_levels));
if (print_old_level) {
ret.push_back("--" + ARG_PRINT_OLD_LEVELS);
}
return ret;
}
void ReduceDBLevelsCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(ReduceDBLevelsCommand::Name());
ret.append(" --" + ARG_NEW_LEVELS + "=<New number of levels>");
ret.append(" [--" + ARG_PRINT_OLD_LEVELS + "]");
ret.append("\n");
}
void ReduceDBLevelsCommand::OverrideBaseCFOptions(
ColumnFamilyOptions* cf_opts) {
LDBCommand::OverrideBaseCFOptions(cf_opts);
cf_opts->num_levels = old_levels_;
cf_opts->max_bytes_for_level_multiplier_additional.resize(cf_opts->num_levels,
1);
// Disable size compaction
cf_opts->max_bytes_for_level_base = 1ULL << 50;
cf_opts->max_bytes_for_level_multiplier = 1;
}
Status ReduceDBLevelsCommand::GetOldNumOfLevels(Options& opt, int* levels) {
ImmutableDBOptions db_options(opt);
EnvOptions soptions;
std::shared_ptr<Cache> tc(
NewLRUCache(opt.max_open_files - 10, opt.table_cache_numshardbits));
const InternalKeyComparator cmp(opt.comparator);
WriteController wc(opt.delayed_write_rate);
WriteBufferManager wb(opt.db_write_buffer_size);
VersionSet versions(db_path_, &db_options, soptions, tc.get(), &wb, &wc,
/*block_cache_tracer=*/nullptr, /*io_tracer=*/nullptr,
/*db_id*/ "", /*db_session_id*/ "");
std::vector<ColumnFamilyDescriptor> dummy;
ColumnFamilyDescriptor dummy_descriptor(kDefaultColumnFamilyName,
2014-02-01 03:44:48 +00:00
ColumnFamilyOptions(opt));
dummy.push_back(dummy_descriptor);
// We rely the VersionSet::Recover to tell us the internal data structures
// in the db. And the Recover() should never do any change
// (like LogAndApply) to the manifest file.
Status st = versions.Recover(dummy);
if (!st.ok()) {
return st;
}
int max = -1;
auto default_cfd = versions.GetColumnFamilySet()->GetDefault();
for (int i = 0; i < default_cfd->NumberLevels(); i++) {
if (default_cfd->current()->storage_info()->NumLevelFiles(i)) {
max = i;
}
}
*levels = max + 1;
return st;
}
void ReduceDBLevelsCommand::DoCommand() {
if (new_levels_ <= 1) {
rocksdb: Replace ASSERT* with EXPECT* in functions that does not return void value Summary: gtest does not use exceptions to fail a unit test by design, and `ASSERT*`s are implemented using `return`. As a consequence we cannot use `ASSERT*` in a function that does not return `void` value ([[ https://code.google.com/p/googletest/wiki/AdvancedGuide#Assertion_Placement | 1]]), and have to fix our existing code. This diff does this in a generic way, with no manual changes. In order to detect all existing `ASSERT*` that are used in functions that doesn't return void value, I change the code to generate compile errors for such cases. In `util/testharness.h` I defined `EXPECT*` assertions, the same way as `ASSERT*`, and redefined `ASSERT*` to return `void`. Then executed: ```lang=bash % USE_CLANG=1 make all -j55 -k 2> build.log % perl -naF: -e 'print "-- -number=".$F[1]." ".$F[0]."\n" if /: error:/' \ build.log | xargs -L 1 perl -spi -e 's/ASSERT/EXPECT/g if $. == $number' % make format ``` After that I reverted back change to `ASSERT*` in `util/testharness.h`. But preserved introduced `EXPECT*`, which is the same as `ASSERT*`. This will be deleted once switched to gtest. This diff is independent and contains manual changes only in `util/testharness.h`. Test Plan: Make sure all tests are passing. ```lang=bash % USE_CLANG=1 make check ``` Reviewers: igor, lgalanis, sdong, yufei.zhu, rven, meyering Reviewed By: meyering Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D33333
2015-03-17 03:52:32 +00:00
exec_state_ =
LDBCommandExecuteResult::Failed("Invalid number of levels.\n");
return;
}
Status st;
PrepareOptions();
int old_level_num = -1;
st = GetOldNumOfLevels(options_, &old_level_num);
if (!st.ok()) {
exec_state_ = LDBCommandExecuteResult::Failed(st.ToString());
return;
}
if (print_old_levels_) {
fprintf(stdout, "The old number of levels in use is %d\n", old_level_num);
}
if (old_level_num <= new_levels_) {
return;
}
old_levels_ = old_level_num;
OpenDB();
if (exec_state_.IsFailed()) {
return;
}
assert(db_ != nullptr);
// Compact the whole DB to put all files to the highest level.
fprintf(stdout, "Compacting the db...\n");
st =
db_->CompactRange(CompactRangeOptions(), GetCfHandle(), nullptr, nullptr);
CloseDB();
if (st.ok()) {
EnvOptions soptions;
st = VersionSet::ReduceNumberOfLevels(db_path_, &options_, soptions,
new_levels_);
}
if (!st.ok()) {
exec_state_ = LDBCommandExecuteResult::Failed(st.ToString());
return;
}
}
const std::string ChangeCompactionStyleCommand::ARG_OLD_COMPACTION_STYLE =
"old_compaction_style";
const std::string ChangeCompactionStyleCommand::ARG_NEW_COMPACTION_STYLE =
"new_compaction_style";
ChangeCompactionStyleCommand::ChangeCompactionStyleCommand(
const std::vector<std::string>& /*params*/,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(options, flags, false,
BuildCmdLineOptions(
{ARG_OLD_COMPACTION_STYLE, ARG_NEW_COMPACTION_STYLE})),
old_compaction_style_(-1),
new_compaction_style_(-1) {
ParseIntOption(option_map_, ARG_OLD_COMPACTION_STYLE, old_compaction_style_,
exec_state_);
if (old_compaction_style_ != kCompactionStyleLevel &&
old_compaction_style_ != kCompactionStyleUniversal) {
exec_state_ = LDBCommandExecuteResult::Failed(
rocksdb: Replace ASSERT* with EXPECT* in functions that does not return void value Summary: gtest does not use exceptions to fail a unit test by design, and `ASSERT*`s are implemented using `return`. As a consequence we cannot use `ASSERT*` in a function that does not return `void` value ([[ https://code.google.com/p/googletest/wiki/AdvancedGuide#Assertion_Placement | 1]]), and have to fix our existing code. This diff does this in a generic way, with no manual changes. In order to detect all existing `ASSERT*` that are used in functions that doesn't return void value, I change the code to generate compile errors for such cases. In `util/testharness.h` I defined `EXPECT*` assertions, the same way as `ASSERT*`, and redefined `ASSERT*` to return `void`. Then executed: ```lang=bash % USE_CLANG=1 make all -j55 -k 2> build.log % perl -naF: -e 'print "-- -number=".$F[1]." ".$F[0]."\n" if /: error:/' \ build.log | xargs -L 1 perl -spi -e 's/ASSERT/EXPECT/g if $. == $number' % make format ``` After that I reverted back change to `ASSERT*` in `util/testharness.h`. But preserved introduced `EXPECT*`, which is the same as `ASSERT*`. This will be deleted once switched to gtest. This diff is independent and contains manual changes only in `util/testharness.h`. Test Plan: Make sure all tests are passing. ```lang=bash % USE_CLANG=1 make check ``` Reviewers: igor, lgalanis, sdong, yufei.zhu, rven, meyering Reviewed By: meyering Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D33333
2015-03-17 03:52:32 +00:00
"Use --" + ARG_OLD_COMPACTION_STYLE + " to specify old compaction " +
"style. Check ldb help for proper compaction style value.\n");
return;
}
ParseIntOption(option_map_, ARG_NEW_COMPACTION_STYLE, new_compaction_style_,
exec_state_);
if (new_compaction_style_ != kCompactionStyleLevel &&
new_compaction_style_ != kCompactionStyleUniversal) {
exec_state_ = LDBCommandExecuteResult::Failed(
rocksdb: Replace ASSERT* with EXPECT* in functions that does not return void value Summary: gtest does not use exceptions to fail a unit test by design, and `ASSERT*`s are implemented using `return`. As a consequence we cannot use `ASSERT*` in a function that does not return `void` value ([[ https://code.google.com/p/googletest/wiki/AdvancedGuide#Assertion_Placement | 1]]), and have to fix our existing code. This diff does this in a generic way, with no manual changes. In order to detect all existing `ASSERT*` that are used in functions that doesn't return void value, I change the code to generate compile errors for such cases. In `util/testharness.h` I defined `EXPECT*` assertions, the same way as `ASSERT*`, and redefined `ASSERT*` to return `void`. Then executed: ```lang=bash % USE_CLANG=1 make all -j55 -k 2> build.log % perl -naF: -e 'print "-- -number=".$F[1]." ".$F[0]."\n" if /: error:/' \ build.log | xargs -L 1 perl -spi -e 's/ASSERT/EXPECT/g if $. == $number' % make format ``` After that I reverted back change to `ASSERT*` in `util/testharness.h`. But preserved introduced `EXPECT*`, which is the same as `ASSERT*`. This will be deleted once switched to gtest. This diff is independent and contains manual changes only in `util/testharness.h`. Test Plan: Make sure all tests are passing. ```lang=bash % USE_CLANG=1 make check ``` Reviewers: igor, lgalanis, sdong, yufei.zhu, rven, meyering Reviewed By: meyering Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D33333
2015-03-17 03:52:32 +00:00
"Use --" + ARG_NEW_COMPACTION_STYLE + " to specify new compaction " +
"style. Check ldb help for proper compaction style value.\n");
return;
}
if (new_compaction_style_ == old_compaction_style_) {
exec_state_ = LDBCommandExecuteResult::Failed(
rocksdb: Replace ASSERT* with EXPECT* in functions that does not return void value Summary: gtest does not use exceptions to fail a unit test by design, and `ASSERT*`s are implemented using `return`. As a consequence we cannot use `ASSERT*` in a function that does not return `void` value ([[ https://code.google.com/p/googletest/wiki/AdvancedGuide#Assertion_Placement | 1]]), and have to fix our existing code. This diff does this in a generic way, with no manual changes. In order to detect all existing `ASSERT*` that are used in functions that doesn't return void value, I change the code to generate compile errors for such cases. In `util/testharness.h` I defined `EXPECT*` assertions, the same way as `ASSERT*`, and redefined `ASSERT*` to return `void`. Then executed: ```lang=bash % USE_CLANG=1 make all -j55 -k 2> build.log % perl -naF: -e 'print "-- -number=".$F[1]." ".$F[0]."\n" if /: error:/' \ build.log | xargs -L 1 perl -spi -e 's/ASSERT/EXPECT/g if $. == $number' % make format ``` After that I reverted back change to `ASSERT*` in `util/testharness.h`. But preserved introduced `EXPECT*`, which is the same as `ASSERT*`. This will be deleted once switched to gtest. This diff is independent and contains manual changes only in `util/testharness.h`. Test Plan: Make sure all tests are passing. ```lang=bash % USE_CLANG=1 make check ``` Reviewers: igor, lgalanis, sdong, yufei.zhu, rven, meyering Reviewed By: meyering Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D33333
2015-03-17 03:52:32 +00:00
"Old compaction style is the same as new compaction style. "
"Nothing to do.\n");
return;
}
if (old_compaction_style_ == kCompactionStyleUniversal &&
new_compaction_style_ == kCompactionStyleLevel) {
exec_state_ = LDBCommandExecuteResult::Failed(
rocksdb: Replace ASSERT* with EXPECT* in functions that does not return void value Summary: gtest does not use exceptions to fail a unit test by design, and `ASSERT*`s are implemented using `return`. As a consequence we cannot use `ASSERT*` in a function that does not return `void` value ([[ https://code.google.com/p/googletest/wiki/AdvancedGuide#Assertion_Placement | 1]]), and have to fix our existing code. This diff does this in a generic way, with no manual changes. In order to detect all existing `ASSERT*` that are used in functions that doesn't return void value, I change the code to generate compile errors for such cases. In `util/testharness.h` I defined `EXPECT*` assertions, the same way as `ASSERT*`, and redefined `ASSERT*` to return `void`. Then executed: ```lang=bash % USE_CLANG=1 make all -j55 -k 2> build.log % perl -naF: -e 'print "-- -number=".$F[1]." ".$F[0]."\n" if /: error:/' \ build.log | xargs -L 1 perl -spi -e 's/ASSERT/EXPECT/g if $. == $number' % make format ``` After that I reverted back change to `ASSERT*` in `util/testharness.h`. But preserved introduced `EXPECT*`, which is the same as `ASSERT*`. This will be deleted once switched to gtest. This diff is independent and contains manual changes only in `util/testharness.h`. Test Plan: Make sure all tests are passing. ```lang=bash % USE_CLANG=1 make check ``` Reviewers: igor, lgalanis, sdong, yufei.zhu, rven, meyering Reviewed By: meyering Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D33333
2015-03-17 03:52:32 +00:00
"Convert from universal compaction to level compaction. "
"Nothing to do.\n");
return;
}
}
void ChangeCompactionStyleCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(ChangeCompactionStyleCommand::Name());
ret.append(" --" + ARG_OLD_COMPACTION_STYLE + "=<Old compaction style: 0 " +
"for level compaction, 1 for universal compaction>");
ret.append(" --" + ARG_NEW_COMPACTION_STYLE + "=<New compaction style: 0 " +
"for level compaction, 1 for universal compaction>");
ret.append("\n");
}
void ChangeCompactionStyleCommand::OverrideBaseCFOptions(
ColumnFamilyOptions* cf_opts) {
LDBCommand::OverrideBaseCFOptions(cf_opts);
if (old_compaction_style_ == kCompactionStyleLevel &&
new_compaction_style_ == kCompactionStyleUniversal) {
// In order to convert from level compaction to universal compaction, we
// need to compact all data into a single file and move it to level 0.
cf_opts->disable_auto_compactions = true;
cf_opts->target_file_size_base = INT_MAX;
cf_opts->target_file_size_multiplier = 1;
cf_opts->max_bytes_for_level_base = INT_MAX;
cf_opts->max_bytes_for_level_multiplier = 1;
}
}
void ChangeCompactionStyleCommand::DoCommand() {
if (!db_) {
assert(GetExecuteState().IsFailed());
return;
}
// print db stats before we have made any change
std::string property;
std::string files_per_level;
for (int i = 0; i < db_->NumberLevels(GetCfHandle()); i++) {
db_->GetProperty(GetCfHandle(),
"rocksdb.num-files-at-level" + std::to_string(i),
&property);
// format print string
char buf[100];
snprintf(buf, sizeof(buf), "%s%s", (i ? "," : ""), property.c_str());
files_per_level += buf;
}
fprintf(stdout, "files per level before compaction: %s\n",
files_per_level.c_str());
// manual compact into a single file and move the file to level 0
CompactRangeOptions compact_options;
compact_options.change_level = true;
compact_options.target_level = 0;
Status s =
db_->CompactRange(compact_options, GetCfHandle(), nullptr, nullptr);
if (!s.ok()) {
std::stringstream oss;
oss << "Compaction failed: " << s.ToString();
exec_state_ = LDBCommandExecuteResult::Failed(oss.str());
return;
}
// verify compaction result
files_per_level = "";
int num_files = 0;
for (int i = 0; i < db_->NumberLevels(GetCfHandle()); i++) {
db_->GetProperty(GetCfHandle(),
"rocksdb.num-files-at-level" + std::to_string(i),
&property);
// format print string
char buf[100];
snprintf(buf, sizeof(buf), "%s%s", (i ? "," : ""), property.c_str());
files_per_level += buf;
num_files = atoi(property.c_str());
// level 0 should have only 1 file
if (i == 0 && num_files != 1) {
rocksdb: Replace ASSERT* with EXPECT* in functions that does not return void value Summary: gtest does not use exceptions to fail a unit test by design, and `ASSERT*`s are implemented using `return`. As a consequence we cannot use `ASSERT*` in a function that does not return `void` value ([[ https://code.google.com/p/googletest/wiki/AdvancedGuide#Assertion_Placement | 1]]), and have to fix our existing code. This diff does this in a generic way, with no manual changes. In order to detect all existing `ASSERT*` that are used in functions that doesn't return void value, I change the code to generate compile errors for such cases. In `util/testharness.h` I defined `EXPECT*` assertions, the same way as `ASSERT*`, and redefined `ASSERT*` to return `void`. Then executed: ```lang=bash % USE_CLANG=1 make all -j55 -k 2> build.log % perl -naF: -e 'print "-- -number=".$F[1]." ".$F[0]."\n" if /: error:/' \ build.log | xargs -L 1 perl -spi -e 's/ASSERT/EXPECT/g if $. == $number' % make format ``` After that I reverted back change to `ASSERT*` in `util/testharness.h`. But preserved introduced `EXPECT*`, which is the same as `ASSERT*`. This will be deleted once switched to gtest. This diff is independent and contains manual changes only in `util/testharness.h`. Test Plan: Make sure all tests are passing. ```lang=bash % USE_CLANG=1 make check ``` Reviewers: igor, lgalanis, sdong, yufei.zhu, rven, meyering Reviewed By: meyering Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D33333
2015-03-17 03:52:32 +00:00
exec_state_ = LDBCommandExecuteResult::Failed(
"Number of db files at "
"level 0 after compaction is " +
std::to_string(num_files) + ", not 1.\n");
return;
}
// other levels should have no file
if (i > 0 && num_files != 0) {
rocksdb: Replace ASSERT* with EXPECT* in functions that does not return void value Summary: gtest does not use exceptions to fail a unit test by design, and `ASSERT*`s are implemented using `return`. As a consequence we cannot use `ASSERT*` in a function that does not return `void` value ([[ https://code.google.com/p/googletest/wiki/AdvancedGuide#Assertion_Placement | 1]]), and have to fix our existing code. This diff does this in a generic way, with no manual changes. In order to detect all existing `ASSERT*` that are used in functions that doesn't return void value, I change the code to generate compile errors for such cases. In `util/testharness.h` I defined `EXPECT*` assertions, the same way as `ASSERT*`, and redefined `ASSERT*` to return `void`. Then executed: ```lang=bash % USE_CLANG=1 make all -j55 -k 2> build.log % perl -naF: -e 'print "-- -number=".$F[1]." ".$F[0]."\n" if /: error:/' \ build.log | xargs -L 1 perl -spi -e 's/ASSERT/EXPECT/g if $. == $number' % make format ``` After that I reverted back change to `ASSERT*` in `util/testharness.h`. But preserved introduced `EXPECT*`, which is the same as `ASSERT*`. This will be deleted once switched to gtest. This diff is independent and contains manual changes only in `util/testharness.h`. Test Plan: Make sure all tests are passing. ```lang=bash % USE_CLANG=1 make check ``` Reviewers: igor, lgalanis, sdong, yufei.zhu, rven, meyering Reviewed By: meyering Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D33333
2015-03-17 03:52:32 +00:00
exec_state_ = LDBCommandExecuteResult::Failed(
"Number of db files at "
"level " +
std::to_string(i) + " after compaction is " +
std::to_string(num_files) + ", not 0.\n");
return;
}
}
fprintf(stdout, "files per level after compaction: %s\n",
files_per_level.c_str());
}
// ----------------------------------------------------------------------------
namespace {
struct StdErrReporter : public log::Reader::Reporter {
void Corruption(size_t /*bytes*/, const Status& s) override {
std::cerr << "Corruption detected in log file " << s.ToString() << "\n";
}
};
class InMemoryHandler : public WriteBatch::Handler {
public:
InMemoryHandler(std::stringstream& row, bool print_values,
bool write_after_commit = false)
: Handler(),
row_(row),
print_values_(print_values),
write_after_commit_(write_after_commit) {}
void commonPutMerge(const Slice& key, const Slice& value) {
std::string k = LDBCommand::StringToHex(key.ToString());
if (print_values_) {
std::string v = LDBCommand::StringToHex(value.ToString());
row_ << k << " : ";
row_ << v << " ";
} else {
row_ << k << " ";
}
}
Status PutCF(uint32_t cf, const Slice& key, const Slice& value) override {
2016-04-29 01:23:29 +00:00
row_ << "PUT(" << cf << ") : ";
commonPutMerge(key, value);
2016-04-29 01:23:29 +00:00
return Status::OK();
}
Wide Column support in ldb (#11754) Summary: wide_columns can now be pretty-printed in the following commands - `./ldb dump_wal` - `./ldb dump` - `./ldb idump` - `./ldb dump_live_files` - `./ldb scan` - `./sst_dump --command=scan` There are opportunities to refactor to reduce some nearly identical code. This PR is initial change to add wide column support in `ldb` and `sst_dump` tool. More PRs to come for the refactor. Pull Request resolved: https://github.com/facebook/rocksdb/pull/11754 Test Plan: **New Tests added** - `WideColumnsHelperTest::DumpWideColumns` - `WideColumnsHelperTest::DumpSliceAsWideColumns` **Changes added to existing tests** - `ExternalSSTFileTest::BasicMixed` added to cover mixed case (This test should have been added in https://github.com/facebook/rocksdb/issues/11688). This test does not verify the ldb or sst_dump output. This test was used to create test SST files having some rows with wide columns and some without and the generated SST files were used to manually test sst_dump_tool. - `createSST()` in `sst_dump_test` now takes `wide_column_one_in` to add wide column value in SST **dump_wal** ``` ./ldb dump_wal --walfile=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/000004.log --print_value --header ``` ``` Sequence,Count,ByteSize,Physical Offset,Key(s) : value 1,1,59,0,PUT_ENTITY(0) : 0x:0x68656C6C6F 0x617474725F6E616D6531:0x666F6F 0x617474725F6E616D6532:0x626172 2,1,34,42,PUT_ENTITY(0) : 0x617474725F6F6E65:0x74776F 0x617474725F7468726565:0x666F7572 3,1,17,7d,PUT(0) : 0x7468697264 : 0x62617A ``` **idump** ``` ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/ idump ``` ``` 'first' seq:1, type:22 => :hello attr_name1:foo attr_name2:bar 'second' seq:2, type:22 => attr_one:two attr_three:four 'third' seq:3, type:1 => baz Internal keys in range: 3 ``` **SST Dump from dump_live_files** ``` ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/ compact ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/ dump_live_files ``` ``` ... ============================== SST Files ============================== /tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/000013.sst level:1 ------------------------------ Process /tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/000013.sst Sst file format: block-based 'first' seq:0, type:22 => :hello attr_name1:foo attr_name2:bar 'second' seq:0, type:22 => attr_one:two attr_three:four 'third' seq:0, type:1 => baz ... ``` **dump** ``` ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/ dump ``` ``` first ==> :hello attr_name1:foo attr_name2:bar second ==> attr_one:two attr_three:four third ==> baz Keys in range: 3 ``` **scan** ``` ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/ scan ``` ``` first : :hello attr_name1:foo attr_name2:bar second : attr_one:two attr_three:four third : baz ``` **sst_dump** ``` ./sst_dump --file=/tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/000013.sst --command=scan ``` ``` options.env is 0x7ff54b296000 Process /tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/000013.sst Sst file format: block-based from [] to [] 'first' seq:0, type:22 => :hello attr_name1:foo attr_name2:bar 'second' seq:0, type:22 => attr_one:two attr_three:four 'third' seq:0, type:1 => baz ``` Reviewed By: ltamasi Differential Revision: D48837999 Pulled By: jaykorean fbshipit-source-id: b0280f0589d2b9716bb9b50530ffcabb397d140f
2023-08-30 19:45:52 +00:00
Status PutEntityCF(uint32_t cf, const Slice& key,
const Slice& value) override {
row_ << "PUT_ENTITY(" << cf << ") : ";
std::string k = LDBCommand::StringToHex(key.ToString());
if (print_values_) {
return WideColumnsHelper::DumpSliceAsWideColumns(value, row_, true);
}
return Status::OK();
}
Status MergeCF(uint32_t cf, const Slice& key, const Slice& value) override {
2016-04-29 01:23:29 +00:00
row_ << "MERGE(" << cf << ") : ";
commonPutMerge(key, value);
2016-04-29 01:23:29 +00:00
return Status::OK();
}
Status MarkNoop(bool) override {
row_ << "NOOP ";
return Status::OK();
}
Status DeleteCF(uint32_t cf, const Slice& key) override {
2016-04-29 01:23:29 +00:00
row_ << "DELETE(" << cf << ") : ";
row_ << LDBCommand::StringToHex(key.ToString()) << " ";
2016-04-29 01:23:29 +00:00
return Status::OK();
}
Status SingleDeleteCF(uint32_t cf, const Slice& key) override {
2016-04-29 01:23:29 +00:00
row_ << "SINGLE_DELETE(" << cf << ") : ";
row_ << LDBCommand::StringToHex(key.ToString()) << " ";
return Status::OK();
}
Status DeleteRangeCF(uint32_t cf, const Slice& begin_key,
const Slice& end_key) override {
row_ << "DELETE_RANGE(" << cf << ") : ";
row_ << LDBCommand::StringToHex(begin_key.ToString()) << " ";
row_ << LDBCommand::StringToHex(end_key.ToString()) << " ";
return Status::OK();
}
Status MarkBeginPrepare(bool unprepare) override {
row_ << "BEGIN_PREPARE(";
row_ << (unprepare ? "true" : "false") << ") ";
2016-04-29 01:23:29 +00:00
return Status::OK();
}
Status MarkEndPrepare(const Slice& xid) override {
2016-04-29 01:23:29 +00:00
row_ << "END_PREPARE(";
row_ << LDBCommand::StringToHex(xid.ToString()) << ") ";
return Status::OK();
}
Status MarkRollback(const Slice& xid) override {
2016-04-29 01:23:29 +00:00
row_ << "ROLLBACK(";
row_ << LDBCommand::StringToHex(xid.ToString()) << ") ";
return Status::OK();
}
Status MarkCommit(const Slice& xid) override {
2016-04-29 01:23:29 +00:00
row_ << "COMMIT(";
row_ << LDBCommand::StringToHex(xid.ToString()) << ") ";
return Status::OK();
}
Status MarkCommitWithTimestamp(const Slice& xid,
const Slice& commit_ts) override {
row_ << "COMMIT_WITH_TIMESTAMP(";
row_ << LDBCommand::StringToHex(xid.ToString()) << ", ";
row_ << LDBCommand::StringToHex(commit_ts.ToString()) << ") ";
return Status::OK();
}
~InMemoryHandler() override {}
protected:
Handler::OptionState WriteAfterCommit() const override {
return write_after_commit_ ? Handler::OptionState::kEnabled
: Handler::OptionState::kDisabled;
}
private:
std::stringstream& row_;
bool print_values_;
bool write_after_commit_;
};
void DumpWalFile(Options options, std::string wal_file, bool print_header,
bool print_values, bool is_write_committed,
LDBCommandExecuteResult* exec_state) {
const auto& fs = options.env->GetFileSystem();
FileOptions soptions(options);
std::unique_ptr<SequentialFileReader> wal_file_reader;
Support read rate-limiting in SequentialFileReader (#9973) Summary: Added rate limiter and read rate-limiting support to SequentialFileReader. I've updated call sites to SequentialFileReader::Read with appropriate IO priority (or left a TODO and specified IO_TOTAL for now). The PR is separated into four commits: the first one added the rate-limiting support, but with some fixes in the unit test since the number of request bytes from rate limiter in SequentialFileReader are not accurate (there is overcharge at EOF). The second commit fixed this by allowing SequentialFileReader to check file size and determine how many bytes are left in the file to read. The third commit added benchmark related code. The fourth commit moved the logic of using file size to avoid overcharging the rate limiter into backup engine (the main user of SequentialFileReader). Pull Request resolved: https://github.com/facebook/rocksdb/pull/9973 Test Plan: - `make check`, backup_engine_test covers usage of SequentialFileReader with rate limiter. - Run db_bench to check if rate limiting is throttling as expected: Verified that reads and writes are together throttled at 2MB/s, and at 0.2MB chunks that are 100ms apart. - Set up: `./db_bench --benchmarks=fillrandom -db=/dev/shm/test_rocksdb` - Benchmark: ``` strace -ttfe read,write ./db_bench --benchmarks=backup -db=/dev/shm/test_rocksdb --backup_rate_limit=2097152 --use_existing_db strace -ttfe read,write ./db_bench --benchmarks=restore -db=/dev/shm/test_rocksdb --restore_rate_limit=2097152 --use_existing_db ``` - db bench on backup and restore to ensure no performance regression. - backup (avg over 50 runs): pre-change: 1.90443e+06 micros/op; post-change: 1.8993e+06 micros/op (improve by 0.2%) - restore (avg over 50 runs): pre-change: 1.79105e+06 micros/op; post-change: 1.78192e+06 micros/op (improve by 0.5%) ``` # Set up ./db_bench --benchmarks=fillrandom -db=/tmp/test_rocksdb -num=10000000 # benchmark TEST_TMPDIR=/tmp/test_rocksdb NUM_RUN=50 for ((j=0;j<$NUM_RUN;j++)) do ./db_bench -db=$TEST_TMPDIR -num=10000000 -benchmarks=backup -use_existing_db | egrep 'backup' # Restore #./db_bench -db=$TEST_TMPDIR -num=10000000 -benchmarks=restore -use_existing_db done > rate_limit.txt && awk -v NUM_RUN=$NUM_RUN '{sum+=$3;sum_sqrt+=$3^2}END{print sum/NUM_RUN, sqrt(sum_sqrt/NUM_RUN-(sum/NUM_RUN)^2)}' rate_limit.txt >> rate_limit_2.txt ``` Reviewed By: hx235 Differential Revision: D36327418 Pulled By: cbi42 fbshipit-source-id: e75d4307cff815945482df5ba630c1e88d064691
2022-05-24 17:28:57 +00:00
Status status = SequentialFileReader::Create(
fs, wal_file, soptions, &wal_file_reader, nullptr /* dbg */,
nullptr /* rate_limiter */);
if (!status.ok()) {
if (exec_state) {
*exec_state = LDBCommandExecuteResult::Failed("Failed to open WAL file " +
status.ToString());
} else {
std::cerr << "Error: Failed to open WAL file " << status.ToString()
<< std::endl;
}
} else {
StdErrReporter reporter;
uint64_t log_number;
FileType type;
// we need the log number, but ParseFilename expects dbname/NNN.log.
std::string sanitized = wal_file;
size_t lastslash = sanitized.rfind('/');
if (lastslash != std::string::npos)
sanitized = sanitized.substr(lastslash + 1);
if (!ParseFileName(sanitized, &log_number, &type)) {
// bogus input, carry on as best we can
log_number = 0;
}
log::Reader reader(options.info_log, std::move(wal_file_reader), &reporter,
Support for single-primary, multi-secondary instances (#4899) Summary: This PR allows RocksDB to run in single-primary, multi-secondary process mode. The writer is a regular RocksDB (e.g. an `DBImpl`) instance playing the role of a primary. Multiple `DBImplSecondary` processes (secondaries) share the same set of SST files, MANIFEST, WAL files with the primary. Secondaries tail the MANIFEST of the primary and apply updates to their own in-memory state of the file system, e.g. `VersionStorageInfo`. This PR has several components: 1. (Originally in #4745). Add a `PathNotFound` subcode to `IOError` to denote the failure when a secondary tries to open a file which has been deleted by the primary. 2. (Similar to #4602). Add `FragmentBufferedReader` to handle partially-read, trailing record at the end of a log from where future read can continue. 3. (Originally in #4710 and #4820). Add implementation of the secondary, i.e. `DBImplSecondary`. 3.1 Tail the primary's MANIFEST during recovery. 3.2 Tail the primary's MANIFEST during normal processing by calling `ReadAndApply`. 3.3 Tailing WAL will be in a future PR. 4. Add an example in 'examples/multi_processes_example.cc' to demonstrate the usage of secondary RocksDB instance in a multi-process setting. Instructions to run the example can be found at the beginning of the source code. Pull Request resolved: https://github.com/facebook/rocksdb/pull/4899 Differential Revision: D14510945 Pulled By: riversand963 fbshipit-source-id: 4ac1c5693e6012ad23f7b4b42d3c374fecbe8886
2019-03-26 23:41:31 +00:00
true /* checksum */, log_number);
std::string scratch;
WriteBatch batch;
Slice record;
std::stringstream row;
if (print_header) {
std::cout << "Sequence,Count,ByteSize,Physical Offset,Key(s)";
if (print_values) {
std::cout << " : value ";
}
std::cout << "\n";
}
while (status.ok() && reader.ReadRecord(&record, &scratch)) {
row.str("");
if (record.size() < WriteBatchInternal::kHeader) {
reporter.Corruption(record.size(),
Status::Corruption("log record too small"));
} else {
status = WriteBatchInternal::SetContents(&batch, record);
if (!status.ok()) {
std::stringstream oss;
oss << "Parsing write batch failed: " << status.ToString();
if (exec_state) {
*exec_state = LDBCommandExecuteResult::Failed(oss.str());
} else {
std::cerr << oss.str() << std::endl;
}
break;
}
row << WriteBatchInternal::Sequence(&batch) << ",";
row << WriteBatchInternal::Count(&batch) << ",";
row << WriteBatchInternal::ByteSize(&batch) << ",";
row << reader.LastRecordOffset() << ",";
InMemoryHandler handler(row, print_values, is_write_committed);
status = batch.Iterate(&handler);
if (!status.ok()) {
if (exec_state) {
std::stringstream oss;
oss << "Print write batch error: " << status.ToString();
*exec_state = LDBCommandExecuteResult::Failed(oss.str());
}
row << "error: " << status.ToString();
break;
}
row << "\n";
}
std::cout << row.str();
}
}
}
} // namespace
const std::string WALDumperCommand::ARG_WAL_FILE = "walfile";
const std::string WALDumperCommand::ARG_WRITE_COMMITTED = "write_committed";
const std::string WALDumperCommand::ARG_PRINT_VALUE = "print_value";
const std::string WALDumperCommand::ARG_PRINT_HEADER = "header";
WALDumperCommand::WALDumperCommand(
const std::vector<std::string>& /*params*/,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(options, flags, true,
BuildCmdLineOptions({ARG_WAL_FILE, ARG_WRITE_COMMITTED,
ARG_PRINT_HEADER, ARG_PRINT_VALUE})),
print_header_(false),
print_values_(false),
is_write_committed_(false) {
wal_file_.clear();
auto itr = options.find(ARG_WAL_FILE);
if (itr != options.end()) {
wal_file_ = itr->second;
}
print_header_ = IsFlagPresent(flags, ARG_PRINT_HEADER);
print_values_ = IsFlagPresent(flags, ARG_PRINT_VALUE);
is_write_committed_ = ParseBooleanOption(options, ARG_WRITE_COMMITTED, true);
if (wal_file_.empty()) {
rocksdb: Replace ASSERT* with EXPECT* in functions that does not return void value Summary: gtest does not use exceptions to fail a unit test by design, and `ASSERT*`s are implemented using `return`. As a consequence we cannot use `ASSERT*` in a function that does not return `void` value ([[ https://code.google.com/p/googletest/wiki/AdvancedGuide#Assertion_Placement | 1]]), and have to fix our existing code. This diff does this in a generic way, with no manual changes. In order to detect all existing `ASSERT*` that are used in functions that doesn't return void value, I change the code to generate compile errors for such cases. In `util/testharness.h` I defined `EXPECT*` assertions, the same way as `ASSERT*`, and redefined `ASSERT*` to return `void`. Then executed: ```lang=bash % USE_CLANG=1 make all -j55 -k 2> build.log % perl -naF: -e 'print "-- -number=".$F[1]." ".$F[0]."\n" if /: error:/' \ build.log | xargs -L 1 perl -spi -e 's/ASSERT/EXPECT/g if $. == $number' % make format ``` After that I reverted back change to `ASSERT*` in `util/testharness.h`. But preserved introduced `EXPECT*`, which is the same as `ASSERT*`. This will be deleted once switched to gtest. This diff is independent and contains manual changes only in `util/testharness.h`. Test Plan: Make sure all tests are passing. ```lang=bash % USE_CLANG=1 make check ``` Reviewers: igor, lgalanis, sdong, yufei.zhu, rven, meyering Reviewed By: meyering Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D33333
2015-03-17 03:52:32 +00:00
exec_state_ = LDBCommandExecuteResult::Failed("Argument " + ARG_WAL_FILE +
" must be specified.");
}
}
void WALDumperCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(WALDumperCommand::Name());
ret.append(" --" + ARG_WAL_FILE + "=<write_ahead_log_file_path>");
ret.append(" [--" + ARG_PRINT_HEADER + "] ");
ret.append(" [--" + ARG_PRINT_VALUE + "] ");
ret.append(" [--" + ARG_WRITE_COMMITTED + "=true|false] ");
ret.append("\n");
}
void WALDumperCommand::DoCommand() {
PrepareOptions();
DumpWalFile(options_, wal_file_, print_header_, print_values_,
is_write_committed_, &exec_state_);
}
// ----------------------------------------------------------------------------
GetCommand::GetCommand(const std::vector<std::string>& params,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(
options, flags, true,
BuildCmdLineOptions({ARG_TTL, ARG_HEX, ARG_KEY_HEX, ARG_VALUE_HEX})) {
if (params.size() != 1) {
exec_state_ = LDBCommandExecuteResult::Failed(
rocksdb: Replace ASSERT* with EXPECT* in functions that does not return void value Summary: gtest does not use exceptions to fail a unit test by design, and `ASSERT*`s are implemented using `return`. As a consequence we cannot use `ASSERT*` in a function that does not return `void` value ([[ https://code.google.com/p/googletest/wiki/AdvancedGuide#Assertion_Placement | 1]]), and have to fix our existing code. This diff does this in a generic way, with no manual changes. In order to detect all existing `ASSERT*` that are used in functions that doesn't return void value, I change the code to generate compile errors for such cases. In `util/testharness.h` I defined `EXPECT*` assertions, the same way as `ASSERT*`, and redefined `ASSERT*` to return `void`. Then executed: ```lang=bash % USE_CLANG=1 make all -j55 -k 2> build.log % perl -naF: -e 'print "-- -number=".$F[1]." ".$F[0]."\n" if /: error:/' \ build.log | xargs -L 1 perl -spi -e 's/ASSERT/EXPECT/g if $. == $number' % make format ``` After that I reverted back change to `ASSERT*` in `util/testharness.h`. But preserved introduced `EXPECT*`, which is the same as `ASSERT*`. This will be deleted once switched to gtest. This diff is independent and contains manual changes only in `util/testharness.h`. Test Plan: Make sure all tests are passing. ```lang=bash % USE_CLANG=1 make check ``` Reviewers: igor, lgalanis, sdong, yufei.zhu, rven, meyering Reviewed By: meyering Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D33333
2015-03-17 03:52:32 +00:00
"<key> must be specified for the get command");
} else {
key_ = params.at(0);
}
if (is_key_hex_) {
key_ = HexToString(key_);
}
}
void GetCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(GetCommand::Name());
ret.append(" <key>");
ret.append(" [--" + ARG_TTL + "]");
ret.append("\n");
}
void GetCommand::DoCommand() {
if (!db_) {
assert(GetExecuteState().IsFailed());
return;
}
std::string value;
Status st = db_->Get(ReadOptions(), GetCfHandle(), key_, &value);
if (st.ok()) {
fprintf(stdout, "%s\n",
(is_value_hex_ ? StringToHex(value) : value).c_str());
} else {
std::stringstream oss;
oss << "Get failed: " << st.ToString();
exec_state_ = LDBCommandExecuteResult::Failed(oss.str());
}
}
// ----------------------------------------------------------------------------
GetEntity and PutEntity Support in ldb (#11796) Summary: - `get_entity` and `put_entity` command support in ldb - Input Format for `put_entity`: `ldb --db=<DB_PATH> put_entity <KEY> <COLUMN_1_NAME>:<COLUMN_1_VALUE> <COLUMN_2_NAME>:<COLUMN_2_VALUE> ...` - Output Format for `get_entity`: `<COLUMN_1_NAME>:<COLUMN_1_VALUE> <COLUMN_2_NAME>:<COLUMN_2_VALUE>` - If `get_entity` is called against non-wide column value (existing behavior), empty key (kDefaultWideColumnName) will be printed, appended by `:<COLUMN_VALUE>` - If `get` is called against wide column value (existing behavior), first column value is printed if the first column name is kDefaultWideColumnName. # Test Checks for `put_entity` and `get_entity` added in `ldb_test.py` ``` ❯ python3 tools/ldb_test.py took 45s at 10:45:44 AM Running testBlobBatchPut... .Running testBlobDump .Running testBlobPut... .Running testBlobStartingLevel... .Running testCheckConsistency... .Running testColumnFamilies... .Running testCountDelimDump... .Running testCountDelimIDump... .Running testDumpLiveFiles... .Running testDumpLoad... Warning: 7 bad lines ignored. .Running testGetProperty... .Running testHexPutGet... .Running testIDumpBasics... .Running testIDumpDecodeBlobIndex... .Running testIngestExternalSst... .Running testInvalidCmdLines... .Running testListColumnFamilies... .Running testListLiveFilesMetadata... .Running testManifestDump... .Running testMiscAdminTask... Compacting the db... Sequence,Count,ByteSize,Physical Offset,Key(s) .Running testSSTDump... .Running testSimpleStringPutGet... .Running testStringBatchPut... .Running testTtlPutGet... .Running testWALDump... . ---------------------------------------------------------------------- Ran 25 tests in 57.742s ``` Manual Test ``` # Invalid format for wide columns ❯ ./ldb --db=/tmp/test_db put_entity x4 x5 Failed: wide column format needs to be <column_name>:<column_value> (did you mean put <key> <value>?) # empty column name (kDefaultWideColumnName) ❯ ./ldb --db=/tmp/test_db put_entity x4 :x5 OK ❯ ./ldb --db=/tmp/test_db get_entity x4 :x5 ❯ ./ldb --db=/tmp/test_db get x4 x5 ❯ ./ldb --db=/tmp/test_db put_entity a1 :z1 b1:c1 b2:f1 OK ❯ ./ldb --db=/tmp/test_db get_entity a1 :z1 b1:c1 b2:f1 # Keeping the existing behavior if `get` was called on wide column values ❯ ./ldb --db=/tmp/test_db get a1 z1 # Scan ❯ ./ldb --db=/tmp/test_db scan a1 ==> b1:c1 b2:f1 x4 ==> x5 x5 ==> cn1:cv1 cn2:cv2 # Scan hex ❯ ./ldb --db=/tmp/test_db scan --hex 0x6131 ==> 0x6231:0x6331 0x6232:0x6631 0x7834 ==> 0x7835 0x7835 ==> 0x636E31:0x637631 0x636E32:0x637632 # More testing with hex values ❯ ./ldb --db=/tmp/test_db get_entity 0x6131 --hex 0x6231:0x6331 0x6232:0x6631 ❯ ./ldb --db=/tmp/test_db get_entity 0x78 --hex Failed: GetEntity failed: NotFound: ❯ ./ldb --db=/tmp/test_db get_entity 0x7834 --hex :0x7835 ❯ ./ldb --db=/tmp/test_db put_entity 0x7834 0x6234:0x6635 --hex OK ❯ ./ldb --db=/tmp/test_db get_entity 0x7834 --hex 0x6234:0x6635 ❯ ./ldb --db=/tmp/test_db get_entity 0x7834 --key_hex b4:f5 ❯ ./ldb --db=/tmp/test_db get_entity x4 --value_hex 0x6234:0x6635 ``` Pull Request resolved: https://github.com/facebook/rocksdb/pull/11796 Reviewed By: jowlyzhang Differential Revision: D48978141 Pulled By: jaykorean fbshipit-source-id: 4f87c222417ed90a6dbf39bd7b0f068b01e68393
2023-09-12 23:32:40 +00:00
GetEntityCommand::GetEntityCommand(
const std::vector<std::string>& params,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(
options, flags, true,
BuildCmdLineOptions({ARG_TTL, ARG_HEX, ARG_KEY_HEX, ARG_VALUE_HEX})) {
if (params.size() != 1) {
exec_state_ = LDBCommandExecuteResult::Failed(
"<key> must be specified for the get_entity command");
} else {
key_ = params.at(0);
}
if (is_key_hex_) {
key_ = HexToString(key_);
}
}
void GetEntityCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(GetEntityCommand::Name());
ret.append(" <key>");
ret.append(" [--" + ARG_TTL + "]");
ret.append("\n");
}
void GetEntityCommand::DoCommand() {
if (!db_) {
assert(GetExecuteState().IsFailed());
return;
}
PinnableWideColumns pinnable_wide_columns;
Status st = db_->GetEntity(ReadOptions(), GetCfHandle(), key_,
&pinnable_wide_columns);
if (st.ok()) {
std::ostringstream oss;
WideColumnsHelper::DumpWideColumns(pinnable_wide_columns.columns(), oss,
is_value_hex_);
fprintf(stdout, "%s\n", oss.str().c_str());
} else {
std::stringstream oss;
oss << "GetEntity failed: " << st.ToString();
exec_state_ = LDBCommandExecuteResult::Failed(oss.str());
}
}
// ----------------------------------------------------------------------------
ApproxSizeCommand::ApproxSizeCommand(
const std::vector<std::string>& /*params*/,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(options, flags, true,
BuildCmdLineOptions(
{ARG_HEX, ARG_KEY_HEX, ARG_VALUE_HEX, ARG_FROM, ARG_TO})) {
if (options.find(ARG_FROM) != options.end()) {
start_key_ = options.find(ARG_FROM)->second;
} else {
rocksdb: Replace ASSERT* with EXPECT* in functions that does not return void value Summary: gtest does not use exceptions to fail a unit test by design, and `ASSERT*`s are implemented using `return`. As a consequence we cannot use `ASSERT*` in a function that does not return `void` value ([[ https://code.google.com/p/googletest/wiki/AdvancedGuide#Assertion_Placement | 1]]), and have to fix our existing code. This diff does this in a generic way, with no manual changes. In order to detect all existing `ASSERT*` that are used in functions that doesn't return void value, I change the code to generate compile errors for such cases. In `util/testharness.h` I defined `EXPECT*` assertions, the same way as `ASSERT*`, and redefined `ASSERT*` to return `void`. Then executed: ```lang=bash % USE_CLANG=1 make all -j55 -k 2> build.log % perl -naF: -e 'print "-- -number=".$F[1]." ".$F[0]."\n" if /: error:/' \ build.log | xargs -L 1 perl -spi -e 's/ASSERT/EXPECT/g if $. == $number' % make format ``` After that I reverted back change to `ASSERT*` in `util/testharness.h`. But preserved introduced `EXPECT*`, which is the same as `ASSERT*`. This will be deleted once switched to gtest. This diff is independent and contains manual changes only in `util/testharness.h`. Test Plan: Make sure all tests are passing. ```lang=bash % USE_CLANG=1 make check ``` Reviewers: igor, lgalanis, sdong, yufei.zhu, rven, meyering Reviewed By: meyering Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D33333
2015-03-17 03:52:32 +00:00
exec_state_ = LDBCommandExecuteResult::Failed(
ARG_FROM + " must be specified for approxsize command");
return;
}
if (options.find(ARG_TO) != options.end()) {
end_key_ = options.find(ARG_TO)->second;
} else {
rocksdb: Replace ASSERT* with EXPECT* in functions that does not return void value Summary: gtest does not use exceptions to fail a unit test by design, and `ASSERT*`s are implemented using `return`. As a consequence we cannot use `ASSERT*` in a function that does not return `void` value ([[ https://code.google.com/p/googletest/wiki/AdvancedGuide#Assertion_Placement | 1]]), and have to fix our existing code. This diff does this in a generic way, with no manual changes. In order to detect all existing `ASSERT*` that are used in functions that doesn't return void value, I change the code to generate compile errors for such cases. In `util/testharness.h` I defined `EXPECT*` assertions, the same way as `ASSERT*`, and redefined `ASSERT*` to return `void`. Then executed: ```lang=bash % USE_CLANG=1 make all -j55 -k 2> build.log % perl -naF: -e 'print "-- -number=".$F[1]." ".$F[0]."\n" if /: error:/' \ build.log | xargs -L 1 perl -spi -e 's/ASSERT/EXPECT/g if $. == $number' % make format ``` After that I reverted back change to `ASSERT*` in `util/testharness.h`. But preserved introduced `EXPECT*`, which is the same as `ASSERT*`. This will be deleted once switched to gtest. This diff is independent and contains manual changes only in `util/testharness.h`. Test Plan: Make sure all tests are passing. ```lang=bash % USE_CLANG=1 make check ``` Reviewers: igor, lgalanis, sdong, yufei.zhu, rven, meyering Reviewed By: meyering Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D33333
2015-03-17 03:52:32 +00:00
exec_state_ = LDBCommandExecuteResult::Failed(
ARG_TO + " must be specified for approxsize command");
return;
}
if (is_key_hex_) {
start_key_ = HexToString(start_key_);
end_key_ = HexToString(end_key_);
}
}
void ApproxSizeCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(ApproxSizeCommand::Name());
ret.append(HelpRangeCmdArgs());
ret.append("\n");
}
void ApproxSizeCommand::DoCommand() {
if (!db_) {
assert(GetExecuteState().IsFailed());
return;
}
Range ranges[1];
ranges[0] = Range(start_key_, end_key_);
uint64_t sizes[1];
Status s = db_->GetApproximateSizes(GetCfHandle(), ranges, 1, sizes);
if (!s.ok()) {
std::stringstream oss;
oss << "ApproximateSize failed: " << s.ToString();
exec_state_ = LDBCommandExecuteResult::Failed(oss.str());
} else {
fprintf(stdout, "%lu\n", (unsigned long)sizes[0]);
}
}
// ----------------------------------------------------------------------------
BatchPutCommand::BatchPutCommand(
const std::vector<std::string>& params,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(options, flags, false,
BuildCmdLineOptions({ARG_TTL, ARG_HEX, ARG_KEY_HEX,
ARG_VALUE_HEX, ARG_CREATE_IF_MISSING})) {
if (params.size() < 2) {
exec_state_ = LDBCommandExecuteResult::Failed(
"At least one <key> <value> pair must be specified batchput.");
} else if (params.size() % 2 != 0) {
exec_state_ = LDBCommandExecuteResult::Failed(
"Equal number of <key>s and <value>s must be specified for batchput.");
} else {
for (size_t i = 0; i < params.size(); i += 2) {
std::string key = params.at(i);
std::string value = params.at(i + 1);
key_values_.push_back(std::pair<std::string, std::string>(
is_key_hex_ ? HexToString(key) : key,
is_value_hex_ ? HexToString(value) : value));
}
}
create_if_missing_ = IsFlagPresent(flags_, ARG_CREATE_IF_MISSING);
}
void BatchPutCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(BatchPutCommand::Name());
ret.append(" <key> <value> [<key> <value>] [..]");
ret.append(" [--" + ARG_CREATE_IF_MISSING + "]");
ret.append(" [--" + ARG_TTL + "]");
ret.append("\n");
}
void BatchPutCommand::DoCommand() {
if (!db_) {
assert(GetExecuteState().IsFailed());
return;
}
WriteBatch batch;
Status st;
std::stringstream oss;
for (std::vector<std::pair<std::string, std::string>>::const_iterator itr =
key_values_.begin();
itr != key_values_.end(); ++itr) {
st = batch.Put(GetCfHandle(), itr->first, itr->second);
if (!st.ok()) {
oss << "Put to write batch failed: " << itr->first << "=>" << itr->second
<< " error: " << st.ToString();
break;
}
}
if (st.ok()) {
st = db_->Write(WriteOptions(), &batch);
if (!st.ok()) {
oss << "Write failed: " << st.ToString();
}
}
if (st.ok()) {
fprintf(stdout, "OK\n");
} else {
exec_state_ = LDBCommandExecuteResult::Failed(oss.str());
}
}
void BatchPutCommand::OverrideBaseOptions() {
LDBCommand::OverrideBaseOptions();
options_.create_if_missing = create_if_missing_;
}
// ----------------------------------------------------------------------------
ScanCommand::ScanCommand(const std::vector<std::string>& /*params*/,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(
options, flags, true,
BuildCmdLineOptions({ARG_TTL, ARG_NO_VALUE, ARG_HEX, ARG_KEY_HEX,
ARG_TO, ARG_VALUE_HEX, ARG_FROM, ARG_TIMESTAMP,
ARG_MAX_KEYS, ARG_TTL_START, ARG_TTL_END})),
2016-03-10 21:34:42 +00:00
start_key_specified_(false),
end_key_specified_(false),
max_keys_scanned_(-1),
no_value_(false) {
auto itr = options.find(ARG_FROM);
if (itr != options.end()) {
start_key_ = itr->second;
if (is_key_hex_) {
start_key_ = HexToString(start_key_);
}
start_key_specified_ = true;
}
itr = options.find(ARG_TO);
if (itr != options.end()) {
end_key_ = itr->second;
if (is_key_hex_) {
end_key_ = HexToString(end_key_);
}
end_key_specified_ = true;
}
std::vector<std::string>::const_iterator vitr =
2016-03-10 21:34:42 +00:00
std::find(flags.begin(), flags.end(), ARG_NO_VALUE);
if (vitr != flags.end()) {
no_value_ = true;
}
itr = options.find(ARG_MAX_KEYS);
if (itr != options.end()) {
try {
#if defined(CYGWIN)
max_keys_scanned_ = strtol(itr->second.c_str(), 0, 10);
#else
max_keys_scanned_ = std::stoi(itr->second);
#endif
} catch (const std::invalid_argument&) {
exec_state_ = LDBCommandExecuteResult::Failed(ARG_MAX_KEYS +
rocksdb: Replace ASSERT* with EXPECT* in functions that does not return void value Summary: gtest does not use exceptions to fail a unit test by design, and `ASSERT*`s are implemented using `return`. As a consequence we cannot use `ASSERT*` in a function that does not return `void` value ([[ https://code.google.com/p/googletest/wiki/AdvancedGuide#Assertion_Placement | 1]]), and have to fix our existing code. This diff does this in a generic way, with no manual changes. In order to detect all existing `ASSERT*` that are used in functions that doesn't return void value, I change the code to generate compile errors for such cases. In `util/testharness.h` I defined `EXPECT*` assertions, the same way as `ASSERT*`, and redefined `ASSERT*` to return `void`. Then executed: ```lang=bash % USE_CLANG=1 make all -j55 -k 2> build.log % perl -naF: -e 'print "-- -number=".$F[1]." ".$F[0]."\n" if /: error:/' \ build.log | xargs -L 1 perl -spi -e 's/ASSERT/EXPECT/g if $. == $number' % make format ``` After that I reverted back change to `ASSERT*` in `util/testharness.h`. But preserved introduced `EXPECT*`, which is the same as `ASSERT*`. This will be deleted once switched to gtest. This diff is independent and contains manual changes only in `util/testharness.h`. Test Plan: Make sure all tests are passing. ```lang=bash % USE_CLANG=1 make check ``` Reviewers: igor, lgalanis, sdong, yufei.zhu, rven, meyering Reviewed By: meyering Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D33333
2015-03-17 03:52:32 +00:00
" has an invalid value");
} catch (const std::out_of_range&) {
rocksdb: Replace ASSERT* with EXPECT* in functions that does not return void value Summary: gtest does not use exceptions to fail a unit test by design, and `ASSERT*`s are implemented using `return`. As a consequence we cannot use `ASSERT*` in a function that does not return `void` value ([[ https://code.google.com/p/googletest/wiki/AdvancedGuide#Assertion_Placement | 1]]), and have to fix our existing code. This diff does this in a generic way, with no manual changes. In order to detect all existing `ASSERT*` that are used in functions that doesn't return void value, I change the code to generate compile errors for such cases. In `util/testharness.h` I defined `EXPECT*` assertions, the same way as `ASSERT*`, and redefined `ASSERT*` to return `void`. Then executed: ```lang=bash % USE_CLANG=1 make all -j55 -k 2> build.log % perl -naF: -e 'print "-- -number=".$F[1]." ".$F[0]."\n" if /: error:/' \ build.log | xargs -L 1 perl -spi -e 's/ASSERT/EXPECT/g if $. == $number' % make format ``` After that I reverted back change to `ASSERT*` in `util/testharness.h`. But preserved introduced `EXPECT*`, which is the same as `ASSERT*`. This will be deleted once switched to gtest. This diff is independent and contains manual changes only in `util/testharness.h`. Test Plan: Make sure all tests are passing. ```lang=bash % USE_CLANG=1 make check ``` Reviewers: igor, lgalanis, sdong, yufei.zhu, rven, meyering Reviewed By: meyering Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D33333
2015-03-17 03:52:32 +00:00
exec_state_ = LDBCommandExecuteResult::Failed(
ARG_MAX_KEYS + " has a value out-of-range");
}
}
}
void ScanCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(ScanCommand::Name());
ret.append(HelpRangeCmdArgs());
ret.append(" [--" + ARG_TTL + "]");
ret.append(" [--" + ARG_TIMESTAMP + "]");
ret.append(" [--" + ARG_MAX_KEYS + "=<N>q] ");
ret.append(" [--" + ARG_TTL_START + "=<N>:- is inclusive]");
ret.append(" [--" + ARG_TTL_END + "=<N>:- is exclusive]");
ret.append(" [--" + ARG_NO_VALUE + "]");
ret.append("\n");
}
void ScanCommand::DoCommand() {
if (!db_) {
assert(GetExecuteState().IsFailed());
return;
}
int num_keys_scanned = 0;
ReadOptions scan_read_opts;
scan_read_opts.total_order_seek = true;
Iterator* it = db_->NewIterator(scan_read_opts, GetCfHandle());
if (start_key_specified_) {
it->Seek(start_key_);
} else {
it->SeekToFirst();
}
int ttl_start;
if (!ParseIntOption(option_map_, ARG_TTL_START, ttl_start, exec_state_)) {
ttl_start = DBWithTTLImpl::kMinTimestamp; // TTL introduction time
}
int ttl_end;
if (!ParseIntOption(option_map_, ARG_TTL_END, ttl_end, exec_state_)) {
ttl_end = DBWithTTLImpl::kMaxTimestamp; // Max time allowed by TTL feature
}
if (ttl_end < ttl_start) {
fprintf(stderr, "Error: End time can't be less than start time\n");
delete it;
return;
}
if (is_db_ttl_ && timestamp_) {
fprintf(stdout, "Scanning key-values from %s to %s\n",
TimeToHumanString(ttl_start).c_str(),
TimeToHumanString(ttl_end).c_str());
}
for (;
it->Valid() && (!end_key_specified_ || it->key().ToString() < end_key_);
it->Next()) {
if (is_db_ttl_) {
TtlIterator* it_ttl = static_cast_with_check<TtlIterator>(it);
Iterator with timestamp (#6255) Summary: Preliminary support for iterator with user timestamp. Current implementation does not consider merge operator and reverse iterator. Auto compaction is also disabled in unit tests. Create an iterator with timestamp. ``` ... read_opts.timestamp = &ts; auto* iter = db->NewIterator(read_opts); // target is key without timestamp. for (iter->Seek(target); iter->Valid(); iter->Next()) {} for (iter->SeekToFirst(); iter->Valid(); iter->Next()) {} delete iter; read_opts.timestamp = &ts1; // lower_bound and upper_bound are without timestamp. read_opts.iterate_lower_bound = &lower_bound; read_opts.iterate_upper_bound = &upper_bound; auto* iter1 = db->NewIterator(read_opts); // Do Seek or SeekToFirst() delete iter1; ``` Test plan (dev server) ``` $make check ``` Simple benchmarking (dev server) 1. The overhead introduced by this PR even when timestamp is disabled. key size: 16 bytes value size: 100 bytes Entries: 1000000 Data reside in main memory, and try to stress iterator. Repeated three times on master and this PR. - Seek without next ``` ./db_bench -db=/dev/shm/rocksdbtest-1000 -benchmarks=fillseq,seekrandom -enable_pipelined_write=false -disable_wal=true -format_version=3 ``` master: 159047.0 ops/sec this PR: 158922.3 ops/sec (2% drop in throughput) - Seek and next 10 times ``` ./db_bench -db=/dev/shm/rocksdbtest-1000 -benchmarks=fillseq,seekrandom -enable_pipelined_write=false -disable_wal=true -format_version=3 -seek_nexts=10 ``` master: 109539.3 ops/sec this PR: 107519.7 ops/sec (2% drop in throughput) Pull Request resolved: https://github.com/facebook/rocksdb/pull/6255 Differential Revision: D19438227 Pulled By: riversand963 fbshipit-source-id: b66b4979486f8474619f4aa6bdd88598870b0746
2020-03-07 00:21:03 +00:00
int rawtime = it_ttl->ttl_timestamp();
if (rawtime < ttl_start || rawtime >= ttl_end) {
continue;
}
if (timestamp_) {
fprintf(stdout, "%s ", TimeToHumanString(rawtime).c_str());
}
}
if (no_value_) {
Minor refactor on LDB command for wide column support and release note (#11777) Summary: As mentioned in https://github.com/facebook/rocksdb/issues/11754 , refactor to clean up some nearly identical logic. This PR changes the existing debugging string format of Scan command as the following. ``` ❯ ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/ scan --hex ``` Before ``` 0x6669727374 : :0x68656C6C6F 0x617474725F6E616D6531:0x666F6F 0x617474725F6E616D6532:0x626172 0x7365636F6E64 : 0x617474725F6F6E65:0x74776F 0x617474725F7468726565:0x666F7572 0x7468697264 : 0x62617A ``` After ``` 0x6669727374 ==> :0x68656C6C6F 0x617474725F6E616D6531:0x666F6F 0x617474725F6E616D6532:0x626172 0x7365636F6E64 ==> 0x617474725F6F6E65:0x74776F 0x617474725F7468726565:0x666F7572 0x7468697264 ==> 0x62617A ``` Pull Request resolved: https://github.com/facebook/rocksdb/pull/11777 Test Plan: ``` ❯ ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/ dump first ==> :hello attr_name1:foo attr_name2:bar second ==> attr_one:two attr_three:four third ==> baz Keys in range: 3 ❯ ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/ scan first ==> :hello attr_name1:foo attr_name2:bar second ==> attr_one:two attr_three:four third ==> baz ❯ ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/ dump --hex 0x6669727374 ==> :0x68656C6C6F 0x617474725F6E616D6531:0x666F6F 0x617474725F6E616D6532:0x626172 0x7365636F6E64 ==> 0x617474725F6F6E65:0x74776F 0x617474725F7468726565:0x666F7572 0x7468697264 ==> 0x62617A Keys in range: 3 ❯ ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/ scan --hex 0x6669727374 ==> :0x68656C6C6F 0x617474725F6E616D6531:0x666F6F 0x617474725F6E616D6532:0x626172 0x7365636F6E64 ==> 0x617474725F6F6E65:0x74776F 0x617474725F7468726565:0x666F7572 0x7468697264 ==> 0x62617A ``` Reviewed By: jowlyzhang Differential Revision: D48876755 Pulled By: jaykorean fbshipit-source-id: b1c608a810fe038999ac528b690d398abf5f21d7
2023-08-31 23:17:03 +00:00
std::string key_str = it->key().ToString();
if (is_key_hex_) {
key_str = StringToHex(key_str);
} else if (ldb_options_.key_formatter) {
key_str = ldb_options_.key_formatter->Format(key_str);
2016-03-10 21:34:42 +00:00
}
Minor refactor on LDB command for wide column support and release note (#11777) Summary: As mentioned in https://github.com/facebook/rocksdb/issues/11754 , refactor to clean up some nearly identical logic. This PR changes the existing debugging string format of Scan command as the following. ``` ❯ ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/ scan --hex ``` Before ``` 0x6669727374 : :0x68656C6C6F 0x617474725F6E616D6531:0x666F6F 0x617474725F6E616D6532:0x626172 0x7365636F6E64 : 0x617474725F6F6E65:0x74776F 0x617474725F7468726565:0x666F7572 0x7468697264 : 0x62617A ``` After ``` 0x6669727374 ==> :0x68656C6C6F 0x617474725F6E616D6531:0x666F6F 0x617474725F6E616D6532:0x626172 0x7365636F6E64 ==> 0x617474725F6F6E65:0x74776F 0x617474725F7468726565:0x666F7572 0x7468697264 ==> 0x62617A ``` Pull Request resolved: https://github.com/facebook/rocksdb/pull/11777 Test Plan: ``` ❯ ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/ dump first ==> :hello attr_name1:foo attr_name2:bar second ==> attr_one:two attr_three:four third ==> baz Keys in range: 3 ❯ ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/ scan first ==> :hello attr_name1:foo attr_name2:bar second ==> attr_one:two attr_three:four third ==> baz ❯ ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/ dump --hex 0x6669727374 ==> :0x68656C6C6F 0x617474725F6E616D6531:0x666F6F 0x617474725F6E616D6532:0x626172 0x7365636F6E64 ==> 0x617474725F6F6E65:0x74776F 0x617474725F7468726565:0x666F7572 0x7468697264 ==> 0x62617A Keys in range: 3 ❯ ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/ scan --hex 0x6669727374 ==> :0x68656C6C6F 0x617474725F6E616D6531:0x666F6F 0x617474725F6E616D6532:0x626172 0x7365636F6E64 ==> 0x617474725F6F6E65:0x74776F 0x617474725F7468726565:0x666F7572 0x7468697264 ==> 0x62617A ``` Reviewed By: jowlyzhang Differential Revision: D48876755 Pulled By: jaykorean fbshipit-source-id: b1c608a810fe038999ac528b690d398abf5f21d7
2023-08-31 23:17:03 +00:00
fprintf(stdout, "%s\n", key_str.c_str());
Wide Column support in ldb (#11754) Summary: wide_columns can now be pretty-printed in the following commands - `./ldb dump_wal` - `./ldb dump` - `./ldb idump` - `./ldb dump_live_files` - `./ldb scan` - `./sst_dump --command=scan` There are opportunities to refactor to reduce some nearly identical code. This PR is initial change to add wide column support in `ldb` and `sst_dump` tool. More PRs to come for the refactor. Pull Request resolved: https://github.com/facebook/rocksdb/pull/11754 Test Plan: **New Tests added** - `WideColumnsHelperTest::DumpWideColumns` - `WideColumnsHelperTest::DumpSliceAsWideColumns` **Changes added to existing tests** - `ExternalSSTFileTest::BasicMixed` added to cover mixed case (This test should have been added in https://github.com/facebook/rocksdb/issues/11688). This test does not verify the ldb or sst_dump output. This test was used to create test SST files having some rows with wide columns and some without and the generated SST files were used to manually test sst_dump_tool. - `createSST()` in `sst_dump_test` now takes `wide_column_one_in` to add wide column value in SST **dump_wal** ``` ./ldb dump_wal --walfile=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/000004.log --print_value --header ``` ``` Sequence,Count,ByteSize,Physical Offset,Key(s) : value 1,1,59,0,PUT_ENTITY(0) : 0x:0x68656C6C6F 0x617474725F6E616D6531:0x666F6F 0x617474725F6E616D6532:0x626172 2,1,34,42,PUT_ENTITY(0) : 0x617474725F6F6E65:0x74776F 0x617474725F7468726565:0x666F7572 3,1,17,7d,PUT(0) : 0x7468697264 : 0x62617A ``` **idump** ``` ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/ idump ``` ``` 'first' seq:1, type:22 => :hello attr_name1:foo attr_name2:bar 'second' seq:2, type:22 => attr_one:two attr_three:four 'third' seq:3, type:1 => baz Internal keys in range: 3 ``` **SST Dump from dump_live_files** ``` ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/ compact ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/ dump_live_files ``` ``` ... ============================== SST Files ============================== /tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/000013.sst level:1 ------------------------------ Process /tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/000013.sst Sst file format: block-based 'first' seq:0, type:22 => :hello attr_name1:foo attr_name2:bar 'second' seq:0, type:22 => attr_one:two attr_three:four 'third' seq:0, type:1 => baz ... ``` **dump** ``` ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/ dump ``` ``` first ==> :hello attr_name1:foo attr_name2:bar second ==> attr_one:two attr_three:four third ==> baz Keys in range: 3 ``` **scan** ``` ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/ scan ``` ``` first : :hello attr_name1:foo attr_name2:bar second : attr_one:two attr_three:four third : baz ``` **sst_dump** ``` ./sst_dump --file=/tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/000013.sst --command=scan ``` ``` options.env is 0x7ff54b296000 Process /tmp/rocksdbtest-226125/db_wide_basic_test_3481961_2308393776696827948/000013.sst Sst file format: block-based from [] to [] 'first' seq:0, type:22 => :hello attr_name1:foo attr_name2:bar 'second' seq:0, type:22 => attr_one:two attr_three:four 'third' seq:0, type:1 => baz ``` Reviewed By: ltamasi Differential Revision: D48837999 Pulled By: jaykorean fbshipit-source-id: b0280f0589d2b9716bb9b50530ffcabb397d140f
2023-08-30 19:45:52 +00:00
} else {
Minor refactor on LDB command for wide column support and release note (#11777) Summary: As mentioned in https://github.com/facebook/rocksdb/issues/11754 , refactor to clean up some nearly identical logic. This PR changes the existing debugging string format of Scan command as the following. ``` ❯ ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/ scan --hex ``` Before ``` 0x6669727374 : :0x68656C6C6F 0x617474725F6E616D6531:0x666F6F 0x617474725F6E616D6532:0x626172 0x7365636F6E64 : 0x617474725F6F6E65:0x74776F 0x617474725F7468726565:0x666F7572 0x7468697264 : 0x62617A ``` After ``` 0x6669727374 ==> :0x68656C6C6F 0x617474725F6E616D6531:0x666F6F 0x617474725F6E616D6532:0x626172 0x7365636F6E64 ==> 0x617474725F6F6E65:0x74776F 0x617474725F7468726565:0x666F7572 0x7468697264 ==> 0x62617A ``` Pull Request resolved: https://github.com/facebook/rocksdb/pull/11777 Test Plan: ``` ❯ ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/ dump first ==> :hello attr_name1:foo attr_name2:bar second ==> attr_one:two attr_three:four third ==> baz Keys in range: 3 ❯ ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/ scan first ==> :hello attr_name1:foo attr_name2:bar second ==> attr_one:two attr_three:four third ==> baz ❯ ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/ dump --hex 0x6669727374 ==> :0x68656C6C6F 0x617474725F6E616D6531:0x666F6F 0x617474725F6E616D6532:0x626172 0x7365636F6E64 ==> 0x617474725F6F6E65:0x74776F 0x617474725F7468726565:0x666F7572 0x7468697264 ==> 0x62617A Keys in range: 3 ❯ ./ldb --db=/tmp/rocksdbtest-226125/db_wide_basic_test_2675429_2308393776696827948/ scan --hex 0x6669727374 ==> :0x68656C6C6F 0x617474725F6E616D6531:0x666F6F 0x617474725F6E616D6532:0x626172 0x7365636F6E64 ==> 0x617474725F6F6E65:0x74776F 0x617474725F7468726565:0x666F7572 0x7468697264 ==> 0x62617A ``` Reviewed By: jowlyzhang Differential Revision: D48876755 Pulled By: jaykorean fbshipit-source-id: b1c608a810fe038999ac528b690d398abf5f21d7
2023-08-31 23:17:03 +00:00
std::string str = is_db_ttl_ ? PrintKeyValue(it->key().ToString(),
it->value().ToString(),
is_key_hex_, is_value_hex_)
: PrintKeyValueOrWideColumns(
it->key(), it->value(), it->columns(),
is_key_hex_, is_value_hex_);
fprintf(stdout, "%s\n", str.c_str());
}
num_keys_scanned++;
if (max_keys_scanned_ >= 0 && num_keys_scanned >= max_keys_scanned_) {
break;
}
}
if (!it->status().ok()) { // Check for any errors found during the scan
exec_state_ = LDBCommandExecuteResult::Failed(it->status().ToString());
}
delete it;
}
// ----------------------------------------------------------------------------
DeleteCommand::DeleteCommand(const std::vector<std::string>& params,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(options, flags, false,
BuildCmdLineOptions({ARG_HEX, ARG_KEY_HEX, ARG_VALUE_HEX})) {
if (params.size() != 1) {
exec_state_ = LDBCommandExecuteResult::Failed(
rocksdb: Replace ASSERT* with EXPECT* in functions that does not return void value Summary: gtest does not use exceptions to fail a unit test by design, and `ASSERT*`s are implemented using `return`. As a consequence we cannot use `ASSERT*` in a function that does not return `void` value ([[ https://code.google.com/p/googletest/wiki/AdvancedGuide#Assertion_Placement | 1]]), and have to fix our existing code. This diff does this in a generic way, with no manual changes. In order to detect all existing `ASSERT*` that are used in functions that doesn't return void value, I change the code to generate compile errors for such cases. In `util/testharness.h` I defined `EXPECT*` assertions, the same way as `ASSERT*`, and redefined `ASSERT*` to return `void`. Then executed: ```lang=bash % USE_CLANG=1 make all -j55 -k 2> build.log % perl -naF: -e 'print "-- -number=".$F[1]." ".$F[0]."\n" if /: error:/' \ build.log | xargs -L 1 perl -spi -e 's/ASSERT/EXPECT/g if $. == $number' % make format ``` After that I reverted back change to `ASSERT*` in `util/testharness.h`. But preserved introduced `EXPECT*`, which is the same as `ASSERT*`. This will be deleted once switched to gtest. This diff is independent and contains manual changes only in `util/testharness.h`. Test Plan: Make sure all tests are passing. ```lang=bash % USE_CLANG=1 make check ``` Reviewers: igor, lgalanis, sdong, yufei.zhu, rven, meyering Reviewed By: meyering Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D33333
2015-03-17 03:52:32 +00:00
"KEY must be specified for the delete command");
} else {
key_ = params.at(0);
if (is_key_hex_) {
key_ = HexToString(key_);
}
}
}
void DeleteCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(DeleteCommand::Name() + " <key>");
ret.append("\n");
}
void DeleteCommand::DoCommand() {
if (!db_) {
assert(GetExecuteState().IsFailed());
return;
}
Status st = db_->Delete(WriteOptions(), GetCfHandle(), key_);
if (st.ok()) {
fprintf(stdout, "OK\n");
} else {
exec_state_ = LDBCommandExecuteResult::Failed(st.ToString());
}
}
SingleDeleteCommand::SingleDeleteCommand(
const std::vector<std::string>& params,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(options, flags, false,
BuildCmdLineOptions({ARG_HEX, ARG_KEY_HEX, ARG_VALUE_HEX})) {
if (params.size() != 1) {
exec_state_ = LDBCommandExecuteResult::Failed(
"KEY must be specified for the single delete command");
} else {
key_ = params.at(0);
if (is_key_hex_) {
key_ = HexToString(key_);
}
}
}
void SingleDeleteCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(SingleDeleteCommand::Name() + " <key>");
ret.append("\n");
}
void SingleDeleteCommand::DoCommand() {
if (!db_) {
assert(GetExecuteState().IsFailed());
return;
}
Status st = db_->SingleDelete(WriteOptions(), GetCfHandle(), key_);
if (st.ok()) {
fprintf(stdout, "OK\n");
} else {
exec_state_ = LDBCommandExecuteResult::Failed(st.ToString());
}
}
DeleteRangeCommand::DeleteRangeCommand(
const std::vector<std::string>& params,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(options, flags, false,
BuildCmdLineOptions({ARG_HEX, ARG_KEY_HEX, ARG_VALUE_HEX})) {
if (params.size() != 2) {
exec_state_ = LDBCommandExecuteResult::Failed(
"begin and end keys must be specified for the delete command");
} else {
begin_key_ = params.at(0);
end_key_ = params.at(1);
if (is_key_hex_) {
begin_key_ = HexToString(begin_key_);
end_key_ = HexToString(end_key_);
}
}
}
void DeleteRangeCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(DeleteRangeCommand::Name() + " <begin key> <end key>");
ret.append("\n");
}
void DeleteRangeCommand::DoCommand() {
if (!db_) {
assert(GetExecuteState().IsFailed());
return;
}
Status st =
db_->DeleteRange(WriteOptions(), GetCfHandle(), begin_key_, end_key_);
if (st.ok()) {
fprintf(stdout, "OK\n");
} else {
exec_state_ = LDBCommandExecuteResult::Failed(st.ToString());
}
}
PutCommand::PutCommand(const std::vector<std::string>& params,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(options, flags, false,
BuildCmdLineOptions({ARG_TTL, ARG_HEX, ARG_KEY_HEX,
ARG_VALUE_HEX, ARG_CREATE_IF_MISSING})) {
if (params.size() != 2) {
exec_state_ = LDBCommandExecuteResult::Failed(
rocksdb: Replace ASSERT* with EXPECT* in functions that does not return void value Summary: gtest does not use exceptions to fail a unit test by design, and `ASSERT*`s are implemented using `return`. As a consequence we cannot use `ASSERT*` in a function that does not return `void` value ([[ https://code.google.com/p/googletest/wiki/AdvancedGuide#Assertion_Placement | 1]]), and have to fix our existing code. This diff does this in a generic way, with no manual changes. In order to detect all existing `ASSERT*` that are used in functions that doesn't return void value, I change the code to generate compile errors for such cases. In `util/testharness.h` I defined `EXPECT*` assertions, the same way as `ASSERT*`, and redefined `ASSERT*` to return `void`. Then executed: ```lang=bash % USE_CLANG=1 make all -j55 -k 2> build.log % perl -naF: -e 'print "-- -number=".$F[1]." ".$F[0]."\n" if /: error:/' \ build.log | xargs -L 1 perl -spi -e 's/ASSERT/EXPECT/g if $. == $number' % make format ``` After that I reverted back change to `ASSERT*` in `util/testharness.h`. But preserved introduced `EXPECT*`, which is the same as `ASSERT*`. This will be deleted once switched to gtest. This diff is independent and contains manual changes only in `util/testharness.h`. Test Plan: Make sure all tests are passing. ```lang=bash % USE_CLANG=1 make check ``` Reviewers: igor, lgalanis, sdong, yufei.zhu, rven, meyering Reviewed By: meyering Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D33333
2015-03-17 03:52:32 +00:00
"<key> and <value> must be specified for the put command");
} else {
key_ = params.at(0);
value_ = params.at(1);
}
if (is_key_hex_) {
key_ = HexToString(key_);
}
if (is_value_hex_) {
value_ = HexToString(value_);
}
create_if_missing_ = IsFlagPresent(flags_, ARG_CREATE_IF_MISSING);
}
void PutCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(PutCommand::Name());
ret.append(" <key> <value>");
ret.append(" [--" + ARG_CREATE_IF_MISSING + "]");
ret.append(" [--" + ARG_TTL + "]");
ret.append("\n");
}
void PutCommand::DoCommand() {
if (!db_) {
assert(GetExecuteState().IsFailed());
return;
}
Status st = db_->Put(WriteOptions(), GetCfHandle(), key_, value_);
if (st.ok()) {
fprintf(stdout, "OK\n");
} else {
exec_state_ = LDBCommandExecuteResult::Failed(st.ToString());
}
}
void PutCommand::OverrideBaseOptions() {
LDBCommand::OverrideBaseOptions();
options_.create_if_missing = create_if_missing_;
}
// ----------------------------------------------------------------------------
GetEntity and PutEntity Support in ldb (#11796) Summary: - `get_entity` and `put_entity` command support in ldb - Input Format for `put_entity`: `ldb --db=<DB_PATH> put_entity <KEY> <COLUMN_1_NAME>:<COLUMN_1_VALUE> <COLUMN_2_NAME>:<COLUMN_2_VALUE> ...` - Output Format for `get_entity`: `<COLUMN_1_NAME>:<COLUMN_1_VALUE> <COLUMN_2_NAME>:<COLUMN_2_VALUE>` - If `get_entity` is called against non-wide column value (existing behavior), empty key (kDefaultWideColumnName) will be printed, appended by `:<COLUMN_VALUE>` - If `get` is called against wide column value (existing behavior), first column value is printed if the first column name is kDefaultWideColumnName. # Test Checks for `put_entity` and `get_entity` added in `ldb_test.py` ``` ❯ python3 tools/ldb_test.py took 45s at 10:45:44 AM Running testBlobBatchPut... .Running testBlobDump .Running testBlobPut... .Running testBlobStartingLevel... .Running testCheckConsistency... .Running testColumnFamilies... .Running testCountDelimDump... .Running testCountDelimIDump... .Running testDumpLiveFiles... .Running testDumpLoad... Warning: 7 bad lines ignored. .Running testGetProperty... .Running testHexPutGet... .Running testIDumpBasics... .Running testIDumpDecodeBlobIndex... .Running testIngestExternalSst... .Running testInvalidCmdLines... .Running testListColumnFamilies... .Running testListLiveFilesMetadata... .Running testManifestDump... .Running testMiscAdminTask... Compacting the db... Sequence,Count,ByteSize,Physical Offset,Key(s) .Running testSSTDump... .Running testSimpleStringPutGet... .Running testStringBatchPut... .Running testTtlPutGet... .Running testWALDump... . ---------------------------------------------------------------------- Ran 25 tests in 57.742s ``` Manual Test ``` # Invalid format for wide columns ❯ ./ldb --db=/tmp/test_db put_entity x4 x5 Failed: wide column format needs to be <column_name>:<column_value> (did you mean put <key> <value>?) # empty column name (kDefaultWideColumnName) ❯ ./ldb --db=/tmp/test_db put_entity x4 :x5 OK ❯ ./ldb --db=/tmp/test_db get_entity x4 :x5 ❯ ./ldb --db=/tmp/test_db get x4 x5 ❯ ./ldb --db=/tmp/test_db put_entity a1 :z1 b1:c1 b2:f1 OK ❯ ./ldb --db=/tmp/test_db get_entity a1 :z1 b1:c1 b2:f1 # Keeping the existing behavior if `get` was called on wide column values ❯ ./ldb --db=/tmp/test_db get a1 z1 # Scan ❯ ./ldb --db=/tmp/test_db scan a1 ==> b1:c1 b2:f1 x4 ==> x5 x5 ==> cn1:cv1 cn2:cv2 # Scan hex ❯ ./ldb --db=/tmp/test_db scan --hex 0x6131 ==> 0x6231:0x6331 0x6232:0x6631 0x7834 ==> 0x7835 0x7835 ==> 0x636E31:0x637631 0x636E32:0x637632 # More testing with hex values ❯ ./ldb --db=/tmp/test_db get_entity 0x6131 --hex 0x6231:0x6331 0x6232:0x6631 ❯ ./ldb --db=/tmp/test_db get_entity 0x78 --hex Failed: GetEntity failed: NotFound: ❯ ./ldb --db=/tmp/test_db get_entity 0x7834 --hex :0x7835 ❯ ./ldb --db=/tmp/test_db put_entity 0x7834 0x6234:0x6635 --hex OK ❯ ./ldb --db=/tmp/test_db get_entity 0x7834 --hex 0x6234:0x6635 ❯ ./ldb --db=/tmp/test_db get_entity 0x7834 --key_hex b4:f5 ❯ ./ldb --db=/tmp/test_db get_entity x4 --value_hex 0x6234:0x6635 ``` Pull Request resolved: https://github.com/facebook/rocksdb/pull/11796 Reviewed By: jowlyzhang Differential Revision: D48978141 Pulled By: jaykorean fbshipit-source-id: 4f87c222417ed90a6dbf39bd7b0f068b01e68393
2023-09-12 23:32:40 +00:00
PutEntityCommand::PutEntityCommand(
const std::vector<std::string>& params,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(options, flags, false,
BuildCmdLineOptions({ARG_TTL, ARG_HEX, ARG_KEY_HEX,
ARG_VALUE_HEX, ARG_CREATE_IF_MISSING})) {
if (params.size() < 2) {
exec_state_ = LDBCommandExecuteResult::Failed(
"<key> and at least one column <column_name>:<column_value> must be "
"specified for the put_entity command");
} else {
auto iter = params.begin();
key_ = *iter;
if (is_key_hex_) {
key_ = HexToString(key_);
}
for (++iter; iter != params.end(); ++iter) {
auto split = StringSplit(*iter, ':');
if (split.size() != 2) {
exec_state_ = LDBCommandExecuteResult::Failed(
"wide column format needs to be <column_name>:<column_value> (did "
"you mean put <key> <value>?)");
return;
}
std::string name(split[0]);
std::string value(split[1]);
if (is_value_hex_) {
name = HexToString(name);
value = HexToString(value);
}
column_names_.push_back(name);
column_values_.push_back(value);
}
}
create_if_missing_ = IsFlagPresent(flags_, ARG_CREATE_IF_MISSING);
}
void PutEntityCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(PutCommand::Name());
ret.append(
" <key> <column1_name>:<column1_value> <column2_name>:<column2_value> "
"<...>");
ret.append(" [--" + ARG_CREATE_IF_MISSING + "]");
ret.append(" [--" + ARG_TTL + "]");
ret.append("\n");
}
void PutEntityCommand::DoCommand() {
if (!db_) {
assert(GetExecuteState().IsFailed());
return;
}
assert(column_names_.size() == column_values_.size());
WideColumns columns;
for (size_t i = 0; i < column_names_.size(); i++) {
WideColumn column(column_names_[i], column_values_[i]);
columns.emplace_back(column);
}
Status st = db_->PutEntity(WriteOptions(), GetCfHandle(), key_, columns);
if (st.ok()) {
fprintf(stdout, "OK\n");
} else {
exec_state_ = LDBCommandExecuteResult::Failed(st.ToString());
}
}
void PutEntityCommand::OverrideBaseOptions() {
LDBCommand::OverrideBaseOptions();
options_.create_if_missing = create_if_missing_;
}
// ----------------------------------------------------------------------------
const char* DBQuerierCommand::HELP_CMD = "help";
const char* DBQuerierCommand::GET_CMD = "get";
const char* DBQuerierCommand::PUT_CMD = "put";
const char* DBQuerierCommand::DELETE_CMD = "delete";
DBQuerierCommand::DBQuerierCommand(
const std::vector<std::string>& /*params*/,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(
options, flags, false,
BuildCmdLineOptions({ARG_TTL, ARG_HEX, ARG_KEY_HEX, ARG_VALUE_HEX})) {
}
void DBQuerierCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(DBQuerierCommand::Name());
ret.append(" [--" + ARG_TTL + "]");
ret.append("\n");
ret.append(
" Starts a REPL shell. Type help for list of available "
"commands.");
ret.append("\n");
}
void DBQuerierCommand::DoCommand() {
if (!db_) {
assert(GetExecuteState().IsFailed());
return;
}
ReadOptions read_options;
WriteOptions write_options;
std::string line;
std::string key;
std::string value;
Status s;
std::stringstream oss;
while (s.ok() && getline(std::cin, line, '\n')) {
// Parse line into std::vector<std::string>
std::vector<std::string> tokens;
size_t pos = 0;
while (true) {
size_t pos2 = line.find(' ', pos);
if (pos2 == std::string::npos) {
break;
}
tokens.push_back(line.substr(pos, pos2 - pos));
pos = pos2 + 1;
}
tokens.push_back(line.substr(pos));
const std::string& cmd = tokens[0];
if (cmd == HELP_CMD) {
fprintf(stdout,
"get <key>\n"
"put <key> <value>\n"
"delete <key>\n");
} else if (cmd == DELETE_CMD && tokens.size() == 2) {
key = (is_key_hex_ ? HexToString(tokens[1]) : tokens[1]);
s = db_->Delete(write_options, GetCfHandle(), Slice(key));
if (s.ok()) {
fprintf(stdout, "Successfully deleted %s\n", tokens[1].c_str());
} else {
oss << "delete " << key << " failed: " << s.ToString();
}
} else if (cmd == PUT_CMD && tokens.size() == 3) {
key = (is_key_hex_ ? HexToString(tokens[1]) : tokens[1]);
value = (is_value_hex_ ? HexToString(tokens[2]) : tokens[2]);
s = db_->Put(write_options, GetCfHandle(), Slice(key), Slice(value));
if (s.ok()) {
fprintf(stdout, "Successfully put %s %s\n", tokens[1].c_str(),
tokens[2].c_str());
} else {
oss << "put " << key << "=>" << value << " failed: " << s.ToString();
}
} else if (cmd == GET_CMD && tokens.size() == 2) {
key = (is_key_hex_ ? HexToString(tokens[1]) : tokens[1]);
s = db_->Get(read_options, GetCfHandle(), Slice(key), &value);
if (s.ok()) {
fprintf(stdout, "%s\n",
PrintKeyValue(key, value, is_key_hex_, is_value_hex_).c_str());
} else {
if (s.IsNotFound()) {
fprintf(stdout, "Not found %s\n", tokens[1].c_str());
} else {
oss << "get " << key << " error: " << s.ToString();
}
}
} else {
fprintf(stdout, "Unknown command %s\n", line.c_str());
}
}
if (!s.ok()) {
exec_state_ = LDBCommandExecuteResult::Failed(oss.str());
}
}
// ----------------------------------------------------------------------------
CheckConsistencyCommand::CheckConsistencyCommand(
const std::vector<std::string>& /*params*/,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(options, flags, true, BuildCmdLineOptions({})) {}
void CheckConsistencyCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(CheckConsistencyCommand::Name());
ret.append("\n");
}
void CheckConsistencyCommand::DoCommand() {
options_.paranoid_checks = true;
options_.num_levels = 64;
OpenDB();
if (exec_state_.IsSucceed() || exec_state_.IsNotStarted()) {
fprintf(stdout, "OK\n");
}
CloseDB();
}
// ----------------------------------------------------------------------------
const std::string CheckPointCommand::ARG_CHECKPOINT_DIR = "checkpoint_dir";
CheckPointCommand::CheckPointCommand(
const std::vector<std::string>& /*params*/,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(options, flags, false /* is_read_only */,
BuildCmdLineOptions({ARG_CHECKPOINT_DIR})) {
auto itr = options.find(ARG_CHECKPOINT_DIR);
if (itr != options.end()) {
checkpoint_dir_ = itr->second;
}
}
void CheckPointCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(CheckPointCommand::Name());
ret.append(" [--" + ARG_CHECKPOINT_DIR + "] ");
ret.append("\n");
}
void CheckPointCommand::DoCommand() {
if (!db_) {
assert(GetExecuteState().IsFailed());
return;
}
Checkpoint* checkpoint;
Status status = Checkpoint::Create(db_, &checkpoint);
status = checkpoint->CreateCheckpoint(checkpoint_dir_);
if (status.ok()) {
fprintf(stdout, "OK\n");
} else {
exec_state_ = LDBCommandExecuteResult::Failed(status.ToString());
}
}
// ----------------------------------------------------------------------------
const std::string RepairCommand::ARG_VERBOSE = "verbose";
RepairCommand::RepairCommand(const std::vector<std::string>& /*params*/,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(options, flags, false, BuildCmdLineOptions({ARG_VERBOSE})) {
verbose_ = IsFlagPresent(flags, ARG_VERBOSE);
}
void RepairCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(RepairCommand::Name());
ret.append(" [--" + ARG_VERBOSE + "]");
ret.append("\n");
}
void RepairCommand::OverrideBaseOptions() {
LDBCommand::OverrideBaseOptions();
auto level = verbose_ ? InfoLogLevel::INFO_LEVEL : InfoLogLevel::WARN_LEVEL;
options_.info_log.reset(new StderrLogger(level));
}
void RepairCommand::DoCommand() {
PrepareOptions();
Status status = RepairDB(db_path_, options_);
if (status.ok()) {
fprintf(stdout, "OK\n");
} else {
exec_state_ = LDBCommandExecuteResult::Failed(status.ToString());
}
}
// ----------------------------------------------------------------------------
const std::string BackupEngineCommand::ARG_NUM_THREADS = "num_threads";
const std::string BackupEngineCommand::ARG_BACKUP_ENV_URI = "backup_env_uri";
const std::string BackupEngineCommand::ARG_BACKUP_FS_URI = "backup_fs_uri";
const std::string BackupEngineCommand::ARG_BACKUP_DIR = "backup_dir";
const std::string BackupEngineCommand::ARG_STDERR_LOG_LEVEL =
"stderr_log_level";
BackupEngineCommand::BackupEngineCommand(
const std::vector<std::string>& /*params*/,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(options, flags, false /* is_read_only */,
BuildCmdLineOptions({ARG_BACKUP_ENV_URI, ARG_BACKUP_FS_URI,
ARG_BACKUP_DIR, ARG_NUM_THREADS,
ARG_STDERR_LOG_LEVEL})),
num_threads_(1) {
auto itr = options.find(ARG_NUM_THREADS);
if (itr != options.end()) {
num_threads_ = std::stoi(itr->second);
}
itr = options.find(ARG_BACKUP_ENV_URI);
if (itr != options.end()) {
backup_env_uri_ = itr->second;
}
itr = options.find(ARG_BACKUP_FS_URI);
if (itr != options.end()) {
backup_fs_uri_ = itr->second;
}
if (!backup_env_uri_.empty() && !backup_fs_uri_.empty()) {
exec_state_ = LDBCommandExecuteResult::Failed(
"you may not specity both --" + ARG_BACKUP_ENV_URI + " and --" +
ARG_BACKUP_FS_URI);
}
itr = options.find(ARG_BACKUP_DIR);
if (itr == options.end()) {
exec_state_ = LDBCommandExecuteResult::Failed("--" + ARG_BACKUP_DIR +
": missing backup directory");
} else {
backup_dir_ = itr->second;
}
itr = options.find(ARG_STDERR_LOG_LEVEL);
if (itr != options.end()) {
int stderr_log_level = std::stoi(itr->second);
if (stderr_log_level < 0 ||
stderr_log_level >= InfoLogLevel::NUM_INFO_LOG_LEVELS) {
exec_state_ = LDBCommandExecuteResult::Failed(
ARG_STDERR_LOG_LEVEL + " must be >= 0 and < " +
std::to_string(InfoLogLevel::NUM_INFO_LOG_LEVELS) + ".");
} else {
logger_.reset(
new StderrLogger(static_cast<InfoLogLevel>(stderr_log_level)));
}
}
}
void BackupEngineCommand::Help(const std::string& name, std::string& ret) {
ret.append(" ");
ret.append(name);
ret.append(" [--" + ARG_BACKUP_ENV_URI + " | --" + ARG_BACKUP_FS_URI + "] ");
ret.append(" [--" + ARG_BACKUP_DIR + "] ");
ret.append(" [--" + ARG_NUM_THREADS + "] ");
ret.append(" [--" + ARG_STDERR_LOG_LEVEL + "=<int (InfoLogLevel)>] ");
ret.append("\n");
}
// ----------------------------------------------------------------------------
BackupCommand::BackupCommand(const std::vector<std::string>& params,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: BackupEngineCommand(params, options, flags) {}
void BackupCommand::Help(std::string& ret) {
BackupEngineCommand::Help(Name(), ret);
}
void BackupCommand::DoCommand() {
BackupEngine* backup_engine;
Status status;
if (!db_) {
assert(GetExecuteState().IsFailed());
return;
}
fprintf(stdout, "open db OK\n");
Env* custom_env = backup_env_guard_.get();
if (custom_env == nullptr) {
Status s =
Env::CreateFromUri(config_options_, backup_env_uri_, backup_fs_uri_,
&custom_env, &backup_env_guard_);
if (!s.ok()) {
exec_state_ = LDBCommandExecuteResult::Failed(s.ToString());
return;
}
}
assert(custom_env != nullptr);
BackupEngineOptions backup_options =
BackupEngineOptions(backup_dir_, custom_env);
backup_options.info_log = logger_.get();
backup_options.max_background_operations = num_threads_;
status = BackupEngine::Open(options_.env, backup_options, &backup_engine);
if (status.ok()) {
fprintf(stdout, "open backup engine OK\n");
} else {
exec_state_ = LDBCommandExecuteResult::Failed(status.ToString());
return;
}
status = backup_engine->CreateNewBackup(db_);
if (status.ok()) {
fprintf(stdout, "create new backup OK\n");
} else {
exec_state_ = LDBCommandExecuteResult::Failed(status.ToString());
return;
}
}
// ----------------------------------------------------------------------------
RestoreCommand::RestoreCommand(
const std::vector<std::string>& params,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: BackupEngineCommand(params, options, flags) {}
void RestoreCommand::Help(std::string& ret) {
BackupEngineCommand::Help(Name(), ret);
}
void RestoreCommand::DoCommand() {
Env* custom_env = backup_env_guard_.get();
if (custom_env == nullptr) {
Status s =
Env::CreateFromUri(config_options_, backup_env_uri_, backup_fs_uri_,
&custom_env, &backup_env_guard_);
if (!s.ok()) {
exec_state_ = LDBCommandExecuteResult::Failed(s.ToString());
return;
}
}
assert(custom_env != nullptr);
std::unique_ptr<BackupEngineReadOnly> restore_engine;
Status status;
{
BackupEngineOptions opts(backup_dir_, custom_env);
opts.info_log = logger_.get();
opts.max_background_operations = num_threads_;
BackupEngineReadOnly* raw_restore_engine_ptr;
status =
BackupEngineReadOnly::Open(options_.env, opts, &raw_restore_engine_ptr);
if (status.ok()) {
restore_engine.reset(raw_restore_engine_ptr);
}
}
if (status.ok()) {
fprintf(stdout, "open restore engine OK\n");
status = restore_engine->RestoreDBFromLatestBackup(db_path_, db_path_);
}
if (status.ok()) {
fprintf(stdout, "restore from backup OK\n");
} else {
exec_state_ = LDBCommandExecuteResult::Failed(status.ToString());
}
}
// ----------------------------------------------------------------------------
namespace {
void DumpSstFile(Options options, std::string filename, bool output_hex,
bool show_properties, bool decode_blob_index,
std::string from_key, std::string to_key) {
if (filename.length() <= 4 ||
filename.rfind(".sst") != filename.length() - 4) {
std::cout << "Invalid sst file name." << std::endl;
return;
}
// no verification
ROCKSDB_NAMESPACE::SstFileDumper dumper(
New backup meta schema, with file temperatures (#9660) Summary: The primary goal of this change is to add support for backing up and restoring (applying on restore) file temperature metadata, without committing to either the DB manifest or the FS reported "current" temperatures being exclusive "source of truth". To achieve this goal, we need to add temperature information to backup metadata, which requires updated backup meta schema. Fortunately I prepared for this in https://github.com/facebook/rocksdb/issues/8069, which began forward compatibility in version 6.19.0 for this kind of schema update. (Previously, backup meta schema was not extensible! Making this schema update public will allow some other "nice to have" features like taking backups with hard links, and avoiding crc32c checksum computation when another checksum is already available.) While schema version 2 is newly public, the default schema version is still 1. Until we change the default, users will need to set to 2 to enable features like temperature data backup+restore. New metadata like temperature information will be ignored with a warning in versions before this change and since 6.19.0. The metadata is considered ignorable because a functioning DB can be restored without it. Some detail: * Some renaming because "future schema" is now just public schema 2. * Initialize some atomics in TestFs (linter reported) * Add temperature hint support to SstFileDumper (used by BackupEngine) Pull Request resolved: https://github.com/facebook/rocksdb/pull/9660 Test Plan: related unit test majorly updated for the new functionality, including some shared testing support for tracking temperatures in a FS. Some other tests and testing hooks into production code also updated for making the backup meta schema change public. Reviewed By: ajkr Differential Revision: D34686968 Pulled By: pdillinger fbshipit-source-id: 3ac1fa3e67ee97ca8a5103d79cc87d872c1d862a
2022-03-18 18:06:17 +00:00
options, filename, Temperature::kUnknown,
2 * 1024 * 1024 /* readahead_size */,
/* verify_checksum */ false, output_hex, decode_blob_index);
Status st = dumper.ReadSequential(true, std::numeric_limits<uint64_t>::max(),
!from_key.empty(), from_key,
!to_key.empty(), to_key);
if (!st.ok()) {
std::cerr << "Error in reading SST file " << filename << st.ToString()
<< std::endl;
return;
}
if (show_properties) {
const ROCKSDB_NAMESPACE::TableProperties* table_properties;
std::shared_ptr<const ROCKSDB_NAMESPACE::TableProperties>
table_properties_from_reader;
st = dumper.ReadTableProperties(&table_properties_from_reader);
if (!st.ok()) {
std::cerr << filename << ": " << st.ToString()
<< ". Try to use initial table properties" << std::endl;
table_properties = dumper.GetInitTableProperties();
} else {
table_properties = table_properties_from_reader.get();
}
if (table_properties != nullptr) {
std::cout << std::endl << "Table Properties:" << std::endl;
std::cout << table_properties->ToString("\n") << std::endl;
}
}
}
void DumpBlobFile(const std::string& filename, bool is_key_hex,
bool is_value_hex, bool dump_uncompressed_blobs) {
using ROCKSDB_NAMESPACE::blob_db::BlobDumpTool;
BlobDumpTool tool;
BlobDumpTool::DisplayType blob_type = is_value_hex
? BlobDumpTool::DisplayType::kHex
: BlobDumpTool::DisplayType::kRaw;
BlobDumpTool::DisplayType show_uncompressed_blob =
dump_uncompressed_blobs ? blob_type : BlobDumpTool::DisplayType::kNone;
BlobDumpTool::DisplayType show_blob =
dump_uncompressed_blobs ? BlobDumpTool::DisplayType::kNone : blob_type;
BlobDumpTool::DisplayType show_key = is_key_hex
? BlobDumpTool::DisplayType::kHex
: BlobDumpTool::DisplayType::kRaw;
Status s = tool.Run(filename, show_key, show_blob, show_uncompressed_blob,
/* show_summary */ true);
if (!s.ok()) {
fprintf(stderr, "Failed: %s\n", s.ToString().c_str());
}
}
} // namespace
DBFileDumperCommand::DBFileDumperCommand(
const std::vector<std::string>& /*params*/,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(options, flags, true,
BuildCmdLineOptions(
{ARG_DECODE_BLOB_INDEX, ARG_DUMP_UNCOMPRESSED_BLOBS})),
decode_blob_index_(IsFlagPresent(flags, ARG_DECODE_BLOB_INDEX)),
dump_uncompressed_blobs_(
IsFlagPresent(flags, ARG_DUMP_UNCOMPRESSED_BLOBS)) {}
void DBFileDumperCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(DBFileDumperCommand::Name());
ret.append(" [--" + ARG_DECODE_BLOB_INDEX + "] ");
ret.append(" [--" + ARG_DUMP_UNCOMPRESSED_BLOBS + "] ");
ret.append("\n");
}
void DBFileDumperCommand::DoCommand() {
if (!db_) {
assert(GetExecuteState().IsFailed());
return;
}
Status s;
// TODO: Use --hex, --key_hex, --value_hex flags consistently for
// dumping manifest file, sst files and blob files.
std::cout << "Manifest File" << std::endl;
std::cout << "==============================" << std::endl;
std::string manifest_filename;
s = ReadFileToString(db_->GetEnv(), CurrentFileName(db_->GetName()),
&manifest_filename);
if (!s.ok() || manifest_filename.empty() ||
manifest_filename.back() != '\n') {
std::cerr << "Error when reading CURRENT file "
<< CurrentFileName(db_->GetName()) << std::endl;
}
// remove the trailing '\n'
manifest_filename.resize(manifest_filename.size() - 1);
std::string manifest_filepath = db_->GetName() + "/" + manifest_filename;
// Correct concatenation of filepath and filename:
// Check that there is no double slashes (or more!) when concatenation
// happens.
manifest_filepath = NormalizePath(manifest_filepath);
std::cout << manifest_filepath << std::endl;
DumpManifestFile(options_, manifest_filepath, false, false, false,
column_families_);
std::cout << std::endl;
std::vector<ColumnFamilyMetaData> column_families;
db_->GetAllColumnFamilyMetaData(&column_families);
for (const auto& column_family : column_families) {
std::cout << "Column family name: " << column_family.name << std::endl;
std::cout << "==============================" << std::endl;
std::cout << std::endl;
std::cout << "SST Files" << std::endl;
std::cout << "==============================" << std::endl;
for (const LevelMetaData& level : column_family.levels) {
for (const SstFileMetaData& sst_file : level.files) {
std::string filename = sst_file.db_path + "/" + sst_file.name;
// Correct concatenation of filepath and filename:
// Check that there is no double slashes (or more!) when concatenation
// happens.
filename = NormalizePath(filename);
std::cout << filename << " level:" << level.level << std::endl;
std::cout << "------------------------------" << std::endl;
DumpSstFile(options_, filename, false, true, decode_blob_index_);
std::cout << std::endl;
}
}
std::cout << "Blob Files" << std::endl;
std::cout << "==============================" << std::endl;
for (const BlobMetaData& blob_file : column_family.blob_files) {
std::string filename =
blob_file.blob_file_path + "/" + blob_file.blob_file_name;
// Correct concatenation of filepath and filename:
// Check that there is no double slashes (or more!) when concatenation
// happens.
filename = NormalizePath(filename);
std::cout << filename << std::endl;
std::cout << "------------------------------" << std::endl;
DumpBlobFile(filename, /* is_key_hex */ false, /* is_value_hex */ false,
dump_uncompressed_blobs_);
std::cout << std::endl;
}
}
std::cout << std::endl;
std::cout << "Write Ahead Log Files" << std::endl;
std::cout << "==============================" << std::endl;
ROCKSDB_NAMESPACE::VectorLogPtr wal_files;
s = db_->GetSortedWalFiles(wal_files);
if (!s.ok()) {
std::cerr << "Error when getting WAL files" << std::endl;
} else {
std::string wal_dir;
if (options_.wal_dir.empty()) {
wal_dir = db_->GetName();
} else {
wal_dir = NormalizePath(options_.wal_dir + "/");
}
for (auto& wal : wal_files) {
// TODO(qyang): option.wal_dir should be passed into ldb command
std::string filename = wal_dir + wal->PathName();
std::cout << filename << std::endl;
// TODO(myabandeh): allow configuring is_write_commited
DumpWalFile(options_, filename, true, true, true /* is_write_commited */,
&exec_state_);
}
}
}
Add list live files metadata (#8446) Summary: Add an argument to ldb to dump live file names, column families, and levels, `list_live_files_metadata`. The output shows all active SST file names, sorted first by column family and then by level. For each level the SST files are sorted alphabetically. Typically, the output looks like this: ``` ./ldb --db=/tmp/test_db list_live_files_metadata Live SST Files: ===== Column Family: default ===== ---------- level 0 ---------- /tmp/test_db/000069.sst ---------- level 1 ---------- /tmp/test_db/000064.sst /tmp/test_db/000065.sst /tmp/test_db/000066.sst /tmp/test_db/000071.sst ---------- level 2 ---------- /tmp/test_db/000038.sst /tmp/test_db/000039.sst /tmp/test_db/000052.sst /tmp/test_db/000067.sst /tmp/test_db/000070.sst ------------------------------ ``` Second, a flag was added `--sort_by_filename`, to change the layout of the output. When this flag is added to the command, the output shows all active SST files sorted by name, in front of which the LSM level and the column family are mentioned. With the same example, the following command would return: ``` ./ldb --db=/tmp/test_db list_live_files_metadata --sort_by_filename Live SST Files: /tmp/test_db/000038.sst : level 2, column family 'default' /tmp/test_db/000039.sst : level 2, column family 'default' /tmp/test_db/000052.sst : level 2, column family 'default' /tmp/test_db/000064.sst : level 1, column family 'default' /tmp/test_db/000065.sst : level 1, column family 'default' /tmp/test_db/000066.sst : level 1, column family 'default' /tmp/test_db/000067.sst : level 2, column family 'default' /tmp/test_db/000069.sst : level 0, column family 'default' /tmp/test_db/000070.sst : level 2, column family 'default' /tmp/test_db/000071.sst : level 1, column family 'default' ------------------------------ ``` Thus, the user can either request to show the files by levels, or sorted by filenames. This PR includes a simple Python unit test that makes sure the file name and level printed out by this new feature matches the one found with an existing feature, `dump_live_file`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8446 Reviewed By: akankshamahajan15 Differential Revision: D29320080 Pulled By: bjlemaire fbshipit-source-id: 01fb7b5637c59010d74c80730a28d815994e7009
2021-06-23 02:06:44 +00:00
const std::string DBLiveFilesMetadataDumperCommand::ARG_SORT_BY_FILENAME =
"sort_by_filename";
DBLiveFilesMetadataDumperCommand::DBLiveFilesMetadataDumperCommand(
const std::vector<std::string>& /*params*/,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(options, flags, true,
BuildCmdLineOptions({ARG_SORT_BY_FILENAME})) {
sort_by_filename_ = IsFlagPresent(flags, ARG_SORT_BY_FILENAME);
}
void DBLiveFilesMetadataDumperCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(DBLiveFilesMetadataDumperCommand::Name());
ret.append(" [--" + ARG_SORT_BY_FILENAME + "] ");
ret.append("\n");
}
void DBLiveFilesMetadataDumperCommand::DoCommand() {
if (!db_) {
assert(GetExecuteState().IsFailed());
return;
}
Status s;
std::vector<ColumnFamilyMetaData> metadata;
db_->GetAllColumnFamilyMetaData(&metadata);
Add list live files metadata (#8446) Summary: Add an argument to ldb to dump live file names, column families, and levels, `list_live_files_metadata`. The output shows all active SST file names, sorted first by column family and then by level. For each level the SST files are sorted alphabetically. Typically, the output looks like this: ``` ./ldb --db=/tmp/test_db list_live_files_metadata Live SST Files: ===== Column Family: default ===== ---------- level 0 ---------- /tmp/test_db/000069.sst ---------- level 1 ---------- /tmp/test_db/000064.sst /tmp/test_db/000065.sst /tmp/test_db/000066.sst /tmp/test_db/000071.sst ---------- level 2 ---------- /tmp/test_db/000038.sst /tmp/test_db/000039.sst /tmp/test_db/000052.sst /tmp/test_db/000067.sst /tmp/test_db/000070.sst ------------------------------ ``` Second, a flag was added `--sort_by_filename`, to change the layout of the output. When this flag is added to the command, the output shows all active SST files sorted by name, in front of which the LSM level and the column family are mentioned. With the same example, the following command would return: ``` ./ldb --db=/tmp/test_db list_live_files_metadata --sort_by_filename Live SST Files: /tmp/test_db/000038.sst : level 2, column family 'default' /tmp/test_db/000039.sst : level 2, column family 'default' /tmp/test_db/000052.sst : level 2, column family 'default' /tmp/test_db/000064.sst : level 1, column family 'default' /tmp/test_db/000065.sst : level 1, column family 'default' /tmp/test_db/000066.sst : level 1, column family 'default' /tmp/test_db/000067.sst : level 2, column family 'default' /tmp/test_db/000069.sst : level 0, column family 'default' /tmp/test_db/000070.sst : level 2, column family 'default' /tmp/test_db/000071.sst : level 1, column family 'default' ------------------------------ ``` Thus, the user can either request to show the files by levels, or sorted by filenames. This PR includes a simple Python unit test that makes sure the file name and level printed out by this new feature matches the one found with an existing feature, `dump_live_file`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8446 Reviewed By: akankshamahajan15 Differential Revision: D29320080 Pulled By: bjlemaire fbshipit-source-id: 01fb7b5637c59010d74c80730a28d815994e7009
2021-06-23 02:06:44 +00:00
if (sort_by_filename_) {
std::cout << "Live SST and Blob Files:" << std::endl;
// tuple of <file path, level, column family name>
std::vector<std::tuple<std::string, int, std::string>> all_files;
for (const auto& column_metadata : metadata) {
// Iterate Levels
const auto& levels = column_metadata.levels;
const std::string& cf = column_metadata.name;
for (const auto& level_metadata : levels) {
// Iterate SST files
const auto& sst_files = level_metadata.files;
int level = level_metadata.level;
for (const auto& sst_metadata : sst_files) {
// The SstFileMetaData.name always starts with "/",
// however SstFileMetaData.db_path is the string provided by
// the user as an input. Therefore we check if we can
// concantenate the two strings directly or if we need to
// drop a possible extra "/" at the end of SstFileMetaData.db_path.
std::string filename =
NormalizePath(sst_metadata.db_path + "/" + sst_metadata.name);
all_files.emplace_back(filename, level, cf);
} // End of for-loop over sst files
} // End of for-loop over levels
const auto& blob_files = column_metadata.blob_files;
for (const auto& blob_metadata : blob_files) {
// The BlobMetaData.blob_file_name always starts with "/",
// however BlobMetaData.blob_file_path is the string provided by
// the user as an input. Therefore we check if we can
// concantenate the two strings directly or if we need to
// drop a possible extra "/" at the end of BlobMetaData.blob_file_path.
std::string filename = NormalizePath(
blob_metadata.blob_file_path + "/" + blob_metadata.blob_file_name);
// Level for blob files is encoded as -1
all_files.emplace_back(filename, -1, cf);
} // End of for-loop over blob files
} // End of for-loop over column metadata
// Sort by filename (i.e. first entry in tuple)
std::sort(all_files.begin(), all_files.end());
for (const auto& item : all_files) {
const std::string& filename = std::get<0>(item);
int level = std::get<1>(item);
const std::string& cf = std::get<2>(item);
if (level == -1) { // Blob File
std::cout << filename << ", column family '" << cf << "'" << std::endl;
} else { // SST file
std::cout << filename << " : level " << level << ", column family '"
<< cf << "'" << std::endl;
}
Add list live files metadata (#8446) Summary: Add an argument to ldb to dump live file names, column families, and levels, `list_live_files_metadata`. The output shows all active SST file names, sorted first by column family and then by level. For each level the SST files are sorted alphabetically. Typically, the output looks like this: ``` ./ldb --db=/tmp/test_db list_live_files_metadata Live SST Files: ===== Column Family: default ===== ---------- level 0 ---------- /tmp/test_db/000069.sst ---------- level 1 ---------- /tmp/test_db/000064.sst /tmp/test_db/000065.sst /tmp/test_db/000066.sst /tmp/test_db/000071.sst ---------- level 2 ---------- /tmp/test_db/000038.sst /tmp/test_db/000039.sst /tmp/test_db/000052.sst /tmp/test_db/000067.sst /tmp/test_db/000070.sst ------------------------------ ``` Second, a flag was added `--sort_by_filename`, to change the layout of the output. When this flag is added to the command, the output shows all active SST files sorted by name, in front of which the LSM level and the column family are mentioned. With the same example, the following command would return: ``` ./ldb --db=/tmp/test_db list_live_files_metadata --sort_by_filename Live SST Files: /tmp/test_db/000038.sst : level 2, column family 'default' /tmp/test_db/000039.sst : level 2, column family 'default' /tmp/test_db/000052.sst : level 2, column family 'default' /tmp/test_db/000064.sst : level 1, column family 'default' /tmp/test_db/000065.sst : level 1, column family 'default' /tmp/test_db/000066.sst : level 1, column family 'default' /tmp/test_db/000067.sst : level 2, column family 'default' /tmp/test_db/000069.sst : level 0, column family 'default' /tmp/test_db/000070.sst : level 2, column family 'default' /tmp/test_db/000071.sst : level 1, column family 'default' ------------------------------ ``` Thus, the user can either request to show the files by levels, or sorted by filenames. This PR includes a simple Python unit test that makes sure the file name and level printed out by this new feature matches the one found with an existing feature, `dump_live_file`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8446 Reviewed By: akankshamahajan15 Differential Revision: D29320080 Pulled By: bjlemaire fbshipit-source-id: 01fb7b5637c59010d74c80730a28d815994e7009
2021-06-23 02:06:44 +00:00
}
} else {
for (const auto& column_metadata : metadata) {
std::cout << "===== Column Family: " << column_metadata.name
Add list live files metadata (#8446) Summary: Add an argument to ldb to dump live file names, column families, and levels, `list_live_files_metadata`. The output shows all active SST file names, sorted first by column family and then by level. For each level the SST files are sorted alphabetically. Typically, the output looks like this: ``` ./ldb --db=/tmp/test_db list_live_files_metadata Live SST Files: ===== Column Family: default ===== ---------- level 0 ---------- /tmp/test_db/000069.sst ---------- level 1 ---------- /tmp/test_db/000064.sst /tmp/test_db/000065.sst /tmp/test_db/000066.sst /tmp/test_db/000071.sst ---------- level 2 ---------- /tmp/test_db/000038.sst /tmp/test_db/000039.sst /tmp/test_db/000052.sst /tmp/test_db/000067.sst /tmp/test_db/000070.sst ------------------------------ ``` Second, a flag was added `--sort_by_filename`, to change the layout of the output. When this flag is added to the command, the output shows all active SST files sorted by name, in front of which the LSM level and the column family are mentioned. With the same example, the following command would return: ``` ./ldb --db=/tmp/test_db list_live_files_metadata --sort_by_filename Live SST Files: /tmp/test_db/000038.sst : level 2, column family 'default' /tmp/test_db/000039.sst : level 2, column family 'default' /tmp/test_db/000052.sst : level 2, column family 'default' /tmp/test_db/000064.sst : level 1, column family 'default' /tmp/test_db/000065.sst : level 1, column family 'default' /tmp/test_db/000066.sst : level 1, column family 'default' /tmp/test_db/000067.sst : level 2, column family 'default' /tmp/test_db/000069.sst : level 0, column family 'default' /tmp/test_db/000070.sst : level 2, column family 'default' /tmp/test_db/000071.sst : level 1, column family 'default' ------------------------------ ``` Thus, the user can either request to show the files by levels, or sorted by filenames. This PR includes a simple Python unit test that makes sure the file name and level printed out by this new feature matches the one found with an existing feature, `dump_live_file`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8446 Reviewed By: akankshamahajan15 Differential Revision: D29320080 Pulled By: bjlemaire fbshipit-source-id: 01fb7b5637c59010d74c80730a28d815994e7009
2021-06-23 02:06:44 +00:00
<< " =====" << std::endl;
std::cout << "Live SST Files:" << std::endl;
// Iterate levels
const auto& levels = column_metadata.levels;
for (const auto& level_metadata : levels) {
std::cout << "---------- level " << level_metadata.level
<< " ----------" << std::endl;
// Iterate SST files
const auto& sst_files = level_metadata.files;
for (const auto& sst_metadata : sst_files) {
// The SstFileMetaData.name always starts with "/",
// however SstFileMetaData.db_path is the string provided by
// the user as an input. Therefore we check if we can
// concantenate the two strings directly or if we need to
// drop a possible extra "/" at the end of SstFileMetaData.db_path.
std::string filename =
NormalizePath(sst_metadata.db_path + "/" + sst_metadata.name);
std::cout << filename << std::endl;
} // End of for-loop over sst files
} // End of for-loop over levels
std::cout << "Live Blob Files:" << std::endl;
const auto& blob_files = column_metadata.blob_files;
for (const auto& blob_metadata : blob_files) {
// The BlobMetaData.blob_file_name always starts with "/",
// however BlobMetaData.blob_file_path is the string provided by
// the user as an input. Therefore we check if we can
// concantenate the two strings directly or if we need to
// drop a possible extra "/" at the end of BlobMetaData.blob_file_path.
std::string filename = NormalizePath(
blob_metadata.blob_file_path + "/" + blob_metadata.blob_file_name);
std::cout << filename << std::endl;
} // End of for-loop over blob files
} // End of for-loop over column metadata
} // End of else ("not sort_by_filename")
Add list live files metadata (#8446) Summary: Add an argument to ldb to dump live file names, column families, and levels, `list_live_files_metadata`. The output shows all active SST file names, sorted first by column family and then by level. For each level the SST files are sorted alphabetically. Typically, the output looks like this: ``` ./ldb --db=/tmp/test_db list_live_files_metadata Live SST Files: ===== Column Family: default ===== ---------- level 0 ---------- /tmp/test_db/000069.sst ---------- level 1 ---------- /tmp/test_db/000064.sst /tmp/test_db/000065.sst /tmp/test_db/000066.sst /tmp/test_db/000071.sst ---------- level 2 ---------- /tmp/test_db/000038.sst /tmp/test_db/000039.sst /tmp/test_db/000052.sst /tmp/test_db/000067.sst /tmp/test_db/000070.sst ------------------------------ ``` Second, a flag was added `--sort_by_filename`, to change the layout of the output. When this flag is added to the command, the output shows all active SST files sorted by name, in front of which the LSM level and the column family are mentioned. With the same example, the following command would return: ``` ./ldb --db=/tmp/test_db list_live_files_metadata --sort_by_filename Live SST Files: /tmp/test_db/000038.sst : level 2, column family 'default' /tmp/test_db/000039.sst : level 2, column family 'default' /tmp/test_db/000052.sst : level 2, column family 'default' /tmp/test_db/000064.sst : level 1, column family 'default' /tmp/test_db/000065.sst : level 1, column family 'default' /tmp/test_db/000066.sst : level 1, column family 'default' /tmp/test_db/000067.sst : level 2, column family 'default' /tmp/test_db/000069.sst : level 0, column family 'default' /tmp/test_db/000070.sst : level 2, column family 'default' /tmp/test_db/000071.sst : level 1, column family 'default' ------------------------------ ``` Thus, the user can either request to show the files by levels, or sorted by filenames. This PR includes a simple Python unit test that makes sure the file name and level printed out by this new feature matches the one found with an existing feature, `dump_live_file`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8446 Reviewed By: akankshamahajan15 Differential Revision: D29320080 Pulled By: bjlemaire fbshipit-source-id: 01fb7b5637c59010d74c80730a28d815994e7009
2021-06-23 02:06:44 +00:00
std::cout << "------------------------------" << std::endl;
}
void WriteExternalSstFilesCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(WriteExternalSstFilesCommand::Name());
ret.append(" <output_sst_path>");
ret.append("\n");
}
WriteExternalSstFilesCommand::WriteExternalSstFilesCommand(
const std::vector<std::string>& params,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(
options, flags, false /* is_read_only */,
BuildCmdLineOptions({ARG_HEX, ARG_KEY_HEX, ARG_VALUE_HEX, ARG_FROM,
ARG_TO, ARG_CREATE_IF_MISSING})) {
create_if_missing_ =
IsFlagPresent(flags, ARG_CREATE_IF_MISSING) ||
ParseBooleanOption(options, ARG_CREATE_IF_MISSING, false);
if (params.size() != 1) {
exec_state_ = LDBCommandExecuteResult::Failed(
"output SST file path must be specified");
} else {
output_sst_path_ = params.at(0);
}
}
void WriteExternalSstFilesCommand::DoCommand() {
if (!db_) {
assert(GetExecuteState().IsFailed());
return;
}
ColumnFamilyHandle* cfh = GetCfHandle();
SstFileWriter sst_file_writer(EnvOptions(), db_->GetOptions(), cfh);
Status status = sst_file_writer.Open(output_sst_path_);
if (!status.ok()) {
exec_state_ = LDBCommandExecuteResult::Failed("failed to open SST file: " +
status.ToString());
return;
}
int bad_lines = 0;
std::string line;
std::ifstream ifs_stdin("/dev/stdin");
std::istream* istream_p = ifs_stdin.is_open() ? &ifs_stdin : &std::cin;
while (getline(*istream_p, line, '\n')) {
std::string key;
std::string value;
if (ParseKeyValue(line, &key, &value, is_key_hex_, is_value_hex_)) {
status = sst_file_writer.Put(key, value);
if (!status.ok()) {
exec_state_ = LDBCommandExecuteResult::Failed(
"failed to write record to file: " + status.ToString());
return;
}
} else if (0 == line.find("Keys in range:")) {
// ignore this line
} else if (0 == line.find("Created bg thread 0x")) {
// ignore this line
} else {
bad_lines++;
}
}
status = sst_file_writer.Finish();
if (!status.ok()) {
exec_state_ = LDBCommandExecuteResult::Failed(
"Failed to finish writing to file: " + status.ToString());
return;
}
if (bad_lines > 0) {
fprintf(stderr, "Warning: %d bad lines ignored.\n", bad_lines);
}
exec_state_ = LDBCommandExecuteResult::Succeed(
"external SST file written to " + output_sst_path_);
}
void WriteExternalSstFilesCommand::OverrideBaseOptions() {
LDBCommand::OverrideBaseOptions();
options_.create_if_missing = create_if_missing_;
}
const std::string IngestExternalSstFilesCommand::ARG_MOVE_FILES = "move_files";
const std::string IngestExternalSstFilesCommand::ARG_SNAPSHOT_CONSISTENCY =
"snapshot_consistency";
const std::string IngestExternalSstFilesCommand::ARG_ALLOW_GLOBAL_SEQNO =
"allow_global_seqno";
const std::string IngestExternalSstFilesCommand::ARG_ALLOW_BLOCKING_FLUSH =
"allow_blocking_flush";
const std::string IngestExternalSstFilesCommand::ARG_INGEST_BEHIND =
"ingest_behind";
const std::string IngestExternalSstFilesCommand::ARG_WRITE_GLOBAL_SEQNO =
"write_global_seqno";
void IngestExternalSstFilesCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(IngestExternalSstFilesCommand::Name());
ret.append(" <input_sst_path>");
ret.append(" [--" + ARG_MOVE_FILES + "] ");
ret.append(" [--" + ARG_SNAPSHOT_CONSISTENCY + "] ");
ret.append(" [--" + ARG_ALLOW_GLOBAL_SEQNO + "] ");
ret.append(" [--" + ARG_ALLOW_BLOCKING_FLUSH + "] ");
ret.append(" [--" + ARG_INGEST_BEHIND + "] ");
ret.append(" [--" + ARG_WRITE_GLOBAL_SEQNO + "] ");
ret.append("\n");
}
IngestExternalSstFilesCommand::IngestExternalSstFilesCommand(
const std::vector<std::string>& params,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(
options, flags, false /* is_read_only */,
BuildCmdLineOptions({ARG_MOVE_FILES, ARG_SNAPSHOT_CONSISTENCY,
ARG_ALLOW_GLOBAL_SEQNO, ARG_CREATE_IF_MISSING,
ARG_ALLOW_BLOCKING_FLUSH, ARG_INGEST_BEHIND,
ARG_WRITE_GLOBAL_SEQNO})),
move_files_(false),
snapshot_consistency_(true),
allow_global_seqno_(true),
allow_blocking_flush_(true),
ingest_behind_(false),
write_global_seqno_(true) {
create_if_missing_ =
IsFlagPresent(flags, ARG_CREATE_IF_MISSING) ||
ParseBooleanOption(options, ARG_CREATE_IF_MISSING, false);
move_files_ = IsFlagPresent(flags, ARG_MOVE_FILES) ||
ParseBooleanOption(options, ARG_MOVE_FILES, false);
snapshot_consistency_ =
IsFlagPresent(flags, ARG_SNAPSHOT_CONSISTENCY) ||
ParseBooleanOption(options, ARG_SNAPSHOT_CONSISTENCY, true);
allow_global_seqno_ =
IsFlagPresent(flags, ARG_ALLOW_GLOBAL_SEQNO) ||
ParseBooleanOption(options, ARG_ALLOW_GLOBAL_SEQNO, true);
allow_blocking_flush_ =
IsFlagPresent(flags, ARG_ALLOW_BLOCKING_FLUSH) ||
ParseBooleanOption(options, ARG_ALLOW_BLOCKING_FLUSH, true);
ingest_behind_ = IsFlagPresent(flags, ARG_INGEST_BEHIND) ||
ParseBooleanOption(options, ARG_INGEST_BEHIND, false);
write_global_seqno_ =
IsFlagPresent(flags, ARG_WRITE_GLOBAL_SEQNO) ||
ParseBooleanOption(options, ARG_WRITE_GLOBAL_SEQNO, true);
if (allow_global_seqno_) {
if (!write_global_seqno_) {
fprintf(stderr,
"Warning: not writing global_seqno to the ingested SST can\n"
"prevent older versions of RocksDB from being able to open it\n");
}
} else {
if (write_global_seqno_) {
exec_state_ = LDBCommandExecuteResult::Failed(
"ldb cannot write global_seqno to the ingested SST when global_seqno "
"is not allowed");
}
}
if (params.size() != 1) {
exec_state_ =
LDBCommandExecuteResult::Failed("input SST path must be specified");
} else {
input_sst_path_ = params.at(0);
}
}
void IngestExternalSstFilesCommand::DoCommand() {
if (!db_) {
assert(GetExecuteState().IsFailed());
return;
}
if (GetExecuteState().IsFailed()) {
return;
}
ColumnFamilyHandle* cfh = GetCfHandle();
IngestExternalFileOptions ifo;
ifo.move_files = move_files_;
ifo.snapshot_consistency = snapshot_consistency_;
ifo.allow_global_seqno = allow_global_seqno_;
ifo.allow_blocking_flush = allow_blocking_flush_;
ifo.ingest_behind = ingest_behind_;
ifo.write_global_seqno = write_global_seqno_;
Status status = db_->IngestExternalFile(cfh, {input_sst_path_}, ifo);
if (!status.ok()) {
exec_state_ = LDBCommandExecuteResult::Failed(
"failed to ingest external SST: " + status.ToString());
} else {
exec_state_ =
LDBCommandExecuteResult::Succeed("external SST files ingested");
}
}
void IngestExternalSstFilesCommand::OverrideBaseOptions() {
LDBCommand::OverrideBaseOptions();
options_.create_if_missing = create_if_missing_;
}
ListFileRangeDeletesCommand::ListFileRangeDeletesCommand(
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(options, flags, true, BuildCmdLineOptions({ARG_MAX_KEYS})) {
auto itr = options.find(ARG_MAX_KEYS);
if (itr != options.end()) {
try {
#if defined(CYGWIN)
max_keys_ = strtol(itr->second.c_str(), 0, 10);
#else
max_keys_ = std::stoi(itr->second);
#endif
} catch (const std::invalid_argument&) {
exec_state_ = LDBCommandExecuteResult::Failed(ARG_MAX_KEYS +
" has an invalid value");
} catch (const std::out_of_range&) {
exec_state_ = LDBCommandExecuteResult::Failed(
ARG_MAX_KEYS + " has a value out-of-range");
}
}
}
void ListFileRangeDeletesCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(ListFileRangeDeletesCommand::Name());
ret.append(" [--" + ARG_MAX_KEYS + "=<N>]");
ret.append(" : print tombstones in SST files.\n");
}
void ListFileRangeDeletesCommand::DoCommand() {
if (!db_) {
assert(GetExecuteState().IsFailed());
return;
}
DBImpl* db_impl = static_cast_with_check<DBImpl>(db_->GetRootDB());
std::string out_str;
Status st =
db_impl->TablesRangeTombstoneSummary(GetCfHandle(), max_keys_, &out_str);
if (st.ok()) {
TEST_SYNC_POINT_CALLBACK(
"ListFileRangeDeletesCommand::DoCommand:BeforePrint", &out_str);
fprintf(stdout, "%s\n", out_str.c_str());
}
}
void UnsafeRemoveSstFileCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(UnsafeRemoveSstFileCommand::Name());
ret.append(" <SST file number>");
ret.append(" ");
ret.append(" MUST NOT be used on a live DB.");
ret.append("\n");
}
UnsafeRemoveSstFileCommand::UnsafeRemoveSstFileCommand(
const std::vector<std::string>& params,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(options, flags, false /* is_read_only */,
BuildCmdLineOptions({})) {
if (params.size() != 1) {
exec_state_ =
LDBCommandExecuteResult::Failed("SST file number must be specified");
} else {
char* endptr = nullptr;
sst_file_number_ = strtoull(params.at(0).c_str(), &endptr, 10 /* base */);
if (endptr == nullptr || *endptr != '\0') {
exec_state_ = LDBCommandExecuteResult::Failed(
"Failed to parse SST file number " + params.at(0));
}
}
}
void UnsafeRemoveSstFileCommand::DoCommand() {
Group rocksdb.sst.read.micros stat by IOActivity flush and compaction (#11288) Summary: **Context:** The existing stat rocksdb.sst.read.micros does not reflect each of compaction and flush cases but aggregate them, which is not so helpful for us to understand IO read behavior of each of them. **Summary** - Update `StopWatch` and `RandomAccessFileReader` to record `rocksdb.sst.read.micros` and `rocksdb.file.{flush/compaction}.read.micros` - Fixed the default histogram in `RandomAccessFileReader` - New field `ReadOptions/IOOptions::io_activity`; Pass `ReadOptions` through paths under db open, flush and compaction to where we can prepare `IOOptions` and pass it to `RandomAccessFileReader` - Use `thread_status_util` for assertion in `DbStressFSWrapper` for continuous testing on we are passing correct `io_activity` under db open, flush and compaction Pull Request resolved: https://github.com/facebook/rocksdb/pull/11288 Test Plan: - **Stress test** - **Db bench 1: rocksdb.sst.read.micros COUNT ≈ sum of rocksdb.file.read.flush.micros's and rocksdb.file.read.compaction.micros's.** (without blob) - May not be exactly the same due to `HistogramStat::Add` only guarantees atomic not accuracy across threads. ``` ./db_bench -db=/dev/shm/testdb/ -statistics=true -benchmarks="fillseq" -key_size=32 -value_size=512 -num=50000 -write_buffer_size=655 -target_file_size_base=655 -disable_auto_compactions=false -compression_type=none -bloom_bits=3 (-use_plain_table=1 -prefix_size=10) ``` ``` // BlockBasedTable rocksdb.sst.read.micros P50 : 2.009374 P95 : 4.968548 P99 : 8.110362 P100 : 43.000000 COUNT : 40456 SUM : 114805 rocksdb.file.read.flush.micros P50 : 1.871841 P95 : 3.872407 P99 : 5.540541 P100 : 43.000000 COUNT : 2250 SUM : 6116 rocksdb.file.read.compaction.micros P50 : 2.023109 P95 : 5.029149 P99 : 8.196910 P100 : 26.000000 COUNT : 38206 SUM : 108689 // PlainTable Does not apply ``` - **Db bench 2: performance** **Read** SETUP: db with 900 files ``` ./db_bench -db=/dev/shm/testdb/ -benchmarks="fillseq" -key_size=32 -value_size=512 -num=50000 -write_buffer_size=655 -disable_auto_compactions=true -target_file_size_base=655 -compression_type=none ```run till convergence ``` ./db_bench -seed=1678564177044286 -use_existing_db=true -db=/dev/shm/testdb -benchmarks=readrandom[-X60] -statistics=true -num=1000000 -disable_auto_compactions=true -compression_type=none -bloom_bits=3 ``` Pre-change `readrandom [AVG 60 runs] : 21568 (± 248) ops/sec` Post-change (no regression, -0.3%) `readrandom [AVG 60 runs] : 21486 (± 236) ops/sec` **Compaction/Flush**run till convergence ``` ./db_bench -db=/dev/shm/testdb2/ -seed=1678564177044286 -benchmarks="fillseq[-X60]" -key_size=32 -value_size=512 -num=50000 -write_buffer_size=655 -disable_auto_compactions=false -target_file_size_base=655 -compression_type=none rocksdb.sst.read.micros COUNT : 33820 rocksdb.sst.read.flush.micros COUNT : 1800 rocksdb.sst.read.compaction.micros COUNT : 32020 ``` Pre-change `fillseq [AVG 46 runs] : 1391 (± 214) ops/sec; 0.7 (± 0.1) MB/sec` Post-change (no regression, ~-0.4%) `fillseq [AVG 46 runs] : 1385 (± 216) ops/sec; 0.7 (± 0.1) MB/sec` Reviewed By: ajkr Differential Revision: D44007011 Pulled By: hx235 fbshipit-source-id: a54c89e4846dfc9a135389edf3f3eedfea257132
2023-04-21 16:07:18 +00:00
// TODO: plumb Env::IOActivity
const ReadOptions read_options;
PrepareOptions();
Add manifest fix-up utility for file temperatures (#9683) Summary: The goal of this change is to allow changes to the "current" (in FileSystem) file temperatures to feed back into DB metadata, so that they can inform decisions and stats reporting. In part because of modular code factoring, it doesn't seem easy to do this automagically, where opening an SST file and observing current Temperature different from expected would trigger a change in metadata and DB manifest write (essentially giving the deep read path access to the write path). It is also difficult to do this while the DB is open because of the limitations of LogAndApply. This change allows updating file temperature metadata on a closed DB using an experimental utility function UpdateManifestForFilesState() or `ldb update_manifest --update_temperatures`. This should suffice for "migration" scenarios where outside tooling has placed or re-arranged DB files into a (different) tiered configuration without going through RocksDB itself (currently, only compaction can change temperature metadata). Some details: * Refactored and added unit test for `ldb unsafe_remove_sst_file` because of shared functionality * Pulled in autovector.h changes from https://github.com/facebook/rocksdb/issues/9546 to fix SuperVersionContext move constructor (related to an older draft of this change) Possible follow-up work: * Support updating manifest with file checksums, such as when a new checksum function is used and want existing DB metadata updated for it. * It's possible that for some repair scenarios, lighter weight than full repair, we might want to support UpdateManifestForFilesState() to modify critical file details like size or checksum using same algorithm. But let's make sure these are differentiated from modifying file details in ways that don't suspect corruption (or require extreme trust). Pull Request resolved: https://github.com/facebook/rocksdb/pull/9683 Test Plan: unit tests added Reviewed By: jay-zhuang Differential Revision: D34798828 Pulled By: pdillinger fbshipit-source-id: cfd83e8fb10761d8c9e7f9c020d68c9106a95554
2022-03-18 23:35:51 +00:00
OfflineManifestWriter w(options_, db_path_);
if (column_families_.empty()) {
column_families_.emplace_back(kDefaultColumnFamilyName, options_);
}
Add manifest fix-up utility for file temperatures (#9683) Summary: The goal of this change is to allow changes to the "current" (in FileSystem) file temperatures to feed back into DB metadata, so that they can inform decisions and stats reporting. In part because of modular code factoring, it doesn't seem easy to do this automagically, where opening an SST file and observing current Temperature different from expected would trigger a change in metadata and DB manifest write (essentially giving the deep read path access to the write path). It is also difficult to do this while the DB is open because of the limitations of LogAndApply. This change allows updating file temperature metadata on a closed DB using an experimental utility function UpdateManifestForFilesState() or `ldb update_manifest --update_temperatures`. This should suffice for "migration" scenarios where outside tooling has placed or re-arranged DB files into a (different) tiered configuration without going through RocksDB itself (currently, only compaction can change temperature metadata). Some details: * Refactored and added unit test for `ldb unsafe_remove_sst_file` because of shared functionality * Pulled in autovector.h changes from https://github.com/facebook/rocksdb/issues/9546 to fix SuperVersionContext move constructor (related to an older draft of this change) Possible follow-up work: * Support updating manifest with file checksums, such as when a new checksum function is used and want existing DB metadata updated for it. * It's possible that for some repair scenarios, lighter weight than full repair, we might want to support UpdateManifestForFilesState() to modify critical file details like size or checksum using same algorithm. But let's make sure these are differentiated from modifying file details in ways that don't suspect corruption (or require extreme trust). Pull Request resolved: https://github.com/facebook/rocksdb/pull/9683 Test Plan: unit tests added Reviewed By: jay-zhuang Differential Revision: D34798828 Pulled By: pdillinger fbshipit-source-id: cfd83e8fb10761d8c9e7f9c020d68c9106a95554
2022-03-18 23:35:51 +00:00
Status s = w.Recover(column_families_);
ColumnFamilyData* cfd = nullptr;
int level = -1;
if (s.ok()) {
FileMetaData* metadata = nullptr;
Add manifest fix-up utility for file temperatures (#9683) Summary: The goal of this change is to allow changes to the "current" (in FileSystem) file temperatures to feed back into DB metadata, so that they can inform decisions and stats reporting. In part because of modular code factoring, it doesn't seem easy to do this automagically, where opening an SST file and observing current Temperature different from expected would trigger a change in metadata and DB manifest write (essentially giving the deep read path access to the write path). It is also difficult to do this while the DB is open because of the limitations of LogAndApply. This change allows updating file temperature metadata on a closed DB using an experimental utility function UpdateManifestForFilesState() or `ldb update_manifest --update_temperatures`. This should suffice for "migration" scenarios where outside tooling has placed or re-arranged DB files into a (different) tiered configuration without going through RocksDB itself (currently, only compaction can change temperature metadata). Some details: * Refactored and added unit test for `ldb unsafe_remove_sst_file` because of shared functionality * Pulled in autovector.h changes from https://github.com/facebook/rocksdb/issues/9546 to fix SuperVersionContext move constructor (related to an older draft of this change) Possible follow-up work: * Support updating manifest with file checksums, such as when a new checksum function is used and want existing DB metadata updated for it. * It's possible that for some repair scenarios, lighter weight than full repair, we might want to support UpdateManifestForFilesState() to modify critical file details like size or checksum using same algorithm. But let's make sure these are differentiated from modifying file details in ways that don't suspect corruption (or require extreme trust). Pull Request resolved: https://github.com/facebook/rocksdb/pull/9683 Test Plan: unit tests added Reviewed By: jay-zhuang Differential Revision: D34798828 Pulled By: pdillinger fbshipit-source-id: cfd83e8fb10761d8c9e7f9c020d68c9106a95554
2022-03-18 23:35:51 +00:00
s = w.Versions().GetMetadataForFile(sst_file_number_, &level, &metadata,
&cfd);
}
if (s.ok()) {
VersionEdit edit;
edit.SetColumnFamily(cfd->GetID());
edit.DeleteFile(level, sst_file_number_);
Sync dir containing CURRENT after RenameFile on CURRENT as much as possible (#10573) Summary: **Context:** Below crash test revealed a bug that directory containing CURRENT file (short for `dir_contains_current_file` below) was not always get synced after a new CURRENT is created and being called with `RenameFile` as part of the creation. This bug exposes a risk that such un-synced directory containing the updated CURRENT can’t survive a host crash (e.g, power loss) hence get corrupted. This then will be followed by a recovery from a corrupted CURRENT that we don't want. The root-cause is that a nullptr `FSDirectory* dir_contains_current_file` sometimes gets passed-down to `SetCurrentFile()` hence in those case `dir_contains_current_file->FSDirectory::FsyncWithDirOptions()` will be skipped (which otherwise will internally call`Env/FS::SyncDic()` ) ``` ./db_stress --acquire_snapshot_one_in=10000 --adaptive_readahead=1 --allow_data_in_errors=True --avoid_unnecessary_blocking_io=0 --backup_max_size=104857600 --backup_one_in=100000 --batch_protection_bytes_per_key=8 --block_size=16384 --bloom_bits=134.8015470676662 --bottommost_compression_type=disable --cache_size=8388608 --checkpoint_one_in=1000000 --checksum_type=kCRC32c --clear_column_family_one_in=0 --compact_files_one_in=1000000 --compact_range_one_in=1000000 --compaction_pri=2 --compaction_ttl=100 --compression_max_dict_buffer_bytes=511 --compression_max_dict_bytes=16384 --compression_type=zstd --compression_use_zstd_dict_trainer=1 --compression_zstd_max_train_bytes=65536 --continuous_verification_interval=0 --data_block_index_type=0 --db=$db --db_write_buffer_size=1048576 --delpercent=5 --delrangepercent=0 --destroy_db_initially=0 --disable_wal=0 --enable_compaction_filter=0 --enable_pipelined_write=1 --expected_values_dir=$exp --fail_if_options_file_error=1 --file_checksum_impl=none --flush_one_in=1000000 --get_current_wal_file_one_in=0 --get_live_files_one_in=1000000 --get_property_one_in=1000000 --get_sorted_wal_files_one_in=0 --index_block_restart_interval=4 --ingest_external_file_one_in=0 --iterpercent=10 --key_len_percent_dist=1,30,69 --level_compaction_dynamic_level_bytes=True --mark_for_compaction_one_file_in=10 --max_background_compactions=20 --max_bytes_for_level_base=10485760 --max_key=10000 --max_key_len=3 --max_manifest_file_size=16384 --max_write_batch_group_size_bytes=64 --max_write_buffer_number=3 --max_write_buffer_size_to_maintain=0 --memtable_prefix_bloom_size_ratio=0.001 --memtable_protection_bytes_per_key=1 --memtable_whole_key_filtering=1 --mmap_read=1 --nooverwritepercent=1 --open_metadata_write_fault_one_in=0 --open_read_fault_one_in=0 --open_write_fault_one_in=0 --ops_per_thread=100000000 --optimize_filters_for_memory=1 --paranoid_file_checks=1 --partition_pinning=2 --pause_background_one_in=1000000 --periodic_compaction_seconds=0 --prefix_size=5 --prefixpercent=5 --prepopulate_block_cache=1 --progress_reports=0 --read_fault_one_in=1000 --readpercent=45 --recycle_log_file_num=0 --reopen=0 --ribbon_starting_level=999 --secondary_cache_fault_one_in=32 --secondary_cache_uri=compressed_secondary_cache://capacity=8388608 --set_options_one_in=10000 --snapshot_hold_ops=100000 --sst_file_manager_bytes_per_sec=0 --sst_file_manager_bytes_per_truncate=0 --subcompactions=3 --sync_fault_injection=1 --target_file_size_base=2097 --target_file_size_multiplier=2 --test_batches_snapshots=1 --top_level_index_pinning=1 --use_full_merge_v1=1 --use_merge=1 --value_size_mult=32 --verify_checksum=1 --verify_checksum_one_in=1000000 --verify_db_one_in=100000 --verify_sst_unique_id_in_manifest=1 --wal_bytes_per_sync=524288 --write_buffer_size=4194 --writepercent=35 ``` ``` stderr: WARNING: prefix_size is non-zero but memtablerep != prefix_hash db_stress: utilities/fault_injection_fs.cc:748: virtual rocksdb::IOStatus rocksdb::FaultInjectionTestFS::RenameFile(const std::string &, const std::string &, const rocksdb::IOOptions &, rocksdb::IODebugContext *): Assertion `tlist.find(tdn.second) == tlist.end()' failed.` ``` **Summary:** The PR ensured the non-test path pass down a non-null dir containing CURRENT (which is by current RocksDB assumption just db_dir) by doing the following: - Renamed `directory_to_fsync` as `dir_contains_current_file` in `SetCurrentFile()` to tighten the association between this directory and CURRENT file - Changed `SetCurrentFile()` API to require `dir_contains_current_file` being passed-in, instead of making it by default nullptr. - Because `SetCurrentFile()`'s `dir_contains_current_file` is passed down from `VersionSet::LogAndApply()` then `VersionSet::ProcessManifestWrites()` (i.e, think about this as a chain of 3 functions related to MANIFEST update), these 2 functions also got refactored to require `dir_contains_current_file` - Updated the non-test-path callers of these 3 functions to obtain and pass in non-nullptr `dir_contains_current_file`, which by current assumption of RocksDB, is the `FSDirectory* db_dir`. - `db_impl` path will obtain `DBImpl::directories_.getDbDir()` while others with no access to such `directories_` are obtained on the fly by creating such object `FileSystem::NewDirectory(..)` and manage it by unique pointers to ensure short life time. Pull Request resolved: https://github.com/facebook/rocksdb/pull/10573 Test Plan: - `make check` - Passed the repro db_stress command - For future improvement, since we currently don't assert dir containing CURRENT to be non-nullptr due to https://github.com/facebook/rocksdb/pull/10573#pullrequestreview-1087698899, there is still chances that future developers mistakenly pass down nullptr dir containing CURRENT thus resulting skipped sync dir and cause the bug again. Therefore a smarter test (e.g, such as quoted from ajkr "(make) unsynced data loss to be dropping files corresponding to unsynced directory entries") is still needed. Reviewed By: ajkr Differential Revision: D39005886 Pulled By: hx235 fbshipit-source-id: 336fb9090d0cfa6ca3dd580db86268007dde7f5a
2022-08-30 00:35:21 +00:00
std::unique_ptr<FSDirectory> db_dir;
s = options_.env->GetFileSystem()->NewDirectory(db_path_, IOOptions(),
&db_dir, nullptr);
if (s.ok()) {
Group rocksdb.sst.read.micros stat by IOActivity flush and compaction (#11288) Summary: **Context:** The existing stat rocksdb.sst.read.micros does not reflect each of compaction and flush cases but aggregate them, which is not so helpful for us to understand IO read behavior of each of them. **Summary** - Update `StopWatch` and `RandomAccessFileReader` to record `rocksdb.sst.read.micros` and `rocksdb.file.{flush/compaction}.read.micros` - Fixed the default histogram in `RandomAccessFileReader` - New field `ReadOptions/IOOptions::io_activity`; Pass `ReadOptions` through paths under db open, flush and compaction to where we can prepare `IOOptions` and pass it to `RandomAccessFileReader` - Use `thread_status_util` for assertion in `DbStressFSWrapper` for continuous testing on we are passing correct `io_activity` under db open, flush and compaction Pull Request resolved: https://github.com/facebook/rocksdb/pull/11288 Test Plan: - **Stress test** - **Db bench 1: rocksdb.sst.read.micros COUNT ≈ sum of rocksdb.file.read.flush.micros's and rocksdb.file.read.compaction.micros's.** (without blob) - May not be exactly the same due to `HistogramStat::Add` only guarantees atomic not accuracy across threads. ``` ./db_bench -db=/dev/shm/testdb/ -statistics=true -benchmarks="fillseq" -key_size=32 -value_size=512 -num=50000 -write_buffer_size=655 -target_file_size_base=655 -disable_auto_compactions=false -compression_type=none -bloom_bits=3 (-use_plain_table=1 -prefix_size=10) ``` ``` // BlockBasedTable rocksdb.sst.read.micros P50 : 2.009374 P95 : 4.968548 P99 : 8.110362 P100 : 43.000000 COUNT : 40456 SUM : 114805 rocksdb.file.read.flush.micros P50 : 1.871841 P95 : 3.872407 P99 : 5.540541 P100 : 43.000000 COUNT : 2250 SUM : 6116 rocksdb.file.read.compaction.micros P50 : 2.023109 P95 : 5.029149 P99 : 8.196910 P100 : 26.000000 COUNT : 38206 SUM : 108689 // PlainTable Does not apply ``` - **Db bench 2: performance** **Read** SETUP: db with 900 files ``` ./db_bench -db=/dev/shm/testdb/ -benchmarks="fillseq" -key_size=32 -value_size=512 -num=50000 -write_buffer_size=655 -disable_auto_compactions=true -target_file_size_base=655 -compression_type=none ```run till convergence ``` ./db_bench -seed=1678564177044286 -use_existing_db=true -db=/dev/shm/testdb -benchmarks=readrandom[-X60] -statistics=true -num=1000000 -disable_auto_compactions=true -compression_type=none -bloom_bits=3 ``` Pre-change `readrandom [AVG 60 runs] : 21568 (± 248) ops/sec` Post-change (no regression, -0.3%) `readrandom [AVG 60 runs] : 21486 (± 236) ops/sec` **Compaction/Flush**run till convergence ``` ./db_bench -db=/dev/shm/testdb2/ -seed=1678564177044286 -benchmarks="fillseq[-X60]" -key_size=32 -value_size=512 -num=50000 -write_buffer_size=655 -disable_auto_compactions=false -target_file_size_base=655 -compression_type=none rocksdb.sst.read.micros COUNT : 33820 rocksdb.sst.read.flush.micros COUNT : 1800 rocksdb.sst.read.compaction.micros COUNT : 32020 ``` Pre-change `fillseq [AVG 46 runs] : 1391 (± 214) ops/sec; 0.7 (± 0.1) MB/sec` Post-change (no regression, ~-0.4%) `fillseq [AVG 46 runs] : 1385 (± 216) ops/sec; 0.7 (± 0.1) MB/sec` Reviewed By: ajkr Differential Revision: D44007011 Pulled By: hx235 fbshipit-source-id: a54c89e4846dfc9a135389edf3f3eedfea257132
2023-04-21 16:07:18 +00:00
s = w.LogAndApply(read_options, cfd, &edit, db_dir.get());
Sync dir containing CURRENT after RenameFile on CURRENT as much as possible (#10573) Summary: **Context:** Below crash test revealed a bug that directory containing CURRENT file (short for `dir_contains_current_file` below) was not always get synced after a new CURRENT is created and being called with `RenameFile` as part of the creation. This bug exposes a risk that such un-synced directory containing the updated CURRENT can’t survive a host crash (e.g, power loss) hence get corrupted. This then will be followed by a recovery from a corrupted CURRENT that we don't want. The root-cause is that a nullptr `FSDirectory* dir_contains_current_file` sometimes gets passed-down to `SetCurrentFile()` hence in those case `dir_contains_current_file->FSDirectory::FsyncWithDirOptions()` will be skipped (which otherwise will internally call`Env/FS::SyncDic()` ) ``` ./db_stress --acquire_snapshot_one_in=10000 --adaptive_readahead=1 --allow_data_in_errors=True --avoid_unnecessary_blocking_io=0 --backup_max_size=104857600 --backup_one_in=100000 --batch_protection_bytes_per_key=8 --block_size=16384 --bloom_bits=134.8015470676662 --bottommost_compression_type=disable --cache_size=8388608 --checkpoint_one_in=1000000 --checksum_type=kCRC32c --clear_column_family_one_in=0 --compact_files_one_in=1000000 --compact_range_one_in=1000000 --compaction_pri=2 --compaction_ttl=100 --compression_max_dict_buffer_bytes=511 --compression_max_dict_bytes=16384 --compression_type=zstd --compression_use_zstd_dict_trainer=1 --compression_zstd_max_train_bytes=65536 --continuous_verification_interval=0 --data_block_index_type=0 --db=$db --db_write_buffer_size=1048576 --delpercent=5 --delrangepercent=0 --destroy_db_initially=0 --disable_wal=0 --enable_compaction_filter=0 --enable_pipelined_write=1 --expected_values_dir=$exp --fail_if_options_file_error=1 --file_checksum_impl=none --flush_one_in=1000000 --get_current_wal_file_one_in=0 --get_live_files_one_in=1000000 --get_property_one_in=1000000 --get_sorted_wal_files_one_in=0 --index_block_restart_interval=4 --ingest_external_file_one_in=0 --iterpercent=10 --key_len_percent_dist=1,30,69 --level_compaction_dynamic_level_bytes=True --mark_for_compaction_one_file_in=10 --max_background_compactions=20 --max_bytes_for_level_base=10485760 --max_key=10000 --max_key_len=3 --max_manifest_file_size=16384 --max_write_batch_group_size_bytes=64 --max_write_buffer_number=3 --max_write_buffer_size_to_maintain=0 --memtable_prefix_bloom_size_ratio=0.001 --memtable_protection_bytes_per_key=1 --memtable_whole_key_filtering=1 --mmap_read=1 --nooverwritepercent=1 --open_metadata_write_fault_one_in=0 --open_read_fault_one_in=0 --open_write_fault_one_in=0 --ops_per_thread=100000000 --optimize_filters_for_memory=1 --paranoid_file_checks=1 --partition_pinning=2 --pause_background_one_in=1000000 --periodic_compaction_seconds=0 --prefix_size=5 --prefixpercent=5 --prepopulate_block_cache=1 --progress_reports=0 --read_fault_one_in=1000 --readpercent=45 --recycle_log_file_num=0 --reopen=0 --ribbon_starting_level=999 --secondary_cache_fault_one_in=32 --secondary_cache_uri=compressed_secondary_cache://capacity=8388608 --set_options_one_in=10000 --snapshot_hold_ops=100000 --sst_file_manager_bytes_per_sec=0 --sst_file_manager_bytes_per_truncate=0 --subcompactions=3 --sync_fault_injection=1 --target_file_size_base=2097 --target_file_size_multiplier=2 --test_batches_snapshots=1 --top_level_index_pinning=1 --use_full_merge_v1=1 --use_merge=1 --value_size_mult=32 --verify_checksum=1 --verify_checksum_one_in=1000000 --verify_db_one_in=100000 --verify_sst_unique_id_in_manifest=1 --wal_bytes_per_sync=524288 --write_buffer_size=4194 --writepercent=35 ``` ``` stderr: WARNING: prefix_size is non-zero but memtablerep != prefix_hash db_stress: utilities/fault_injection_fs.cc:748: virtual rocksdb::IOStatus rocksdb::FaultInjectionTestFS::RenameFile(const std::string &, const std::string &, const rocksdb::IOOptions &, rocksdb::IODebugContext *): Assertion `tlist.find(tdn.second) == tlist.end()' failed.` ``` **Summary:** The PR ensured the non-test path pass down a non-null dir containing CURRENT (which is by current RocksDB assumption just db_dir) by doing the following: - Renamed `directory_to_fsync` as `dir_contains_current_file` in `SetCurrentFile()` to tighten the association between this directory and CURRENT file - Changed `SetCurrentFile()` API to require `dir_contains_current_file` being passed-in, instead of making it by default nullptr. - Because `SetCurrentFile()`'s `dir_contains_current_file` is passed down from `VersionSet::LogAndApply()` then `VersionSet::ProcessManifestWrites()` (i.e, think about this as a chain of 3 functions related to MANIFEST update), these 2 functions also got refactored to require `dir_contains_current_file` - Updated the non-test-path callers of these 3 functions to obtain and pass in non-nullptr `dir_contains_current_file`, which by current assumption of RocksDB, is the `FSDirectory* db_dir`. - `db_impl` path will obtain `DBImpl::directories_.getDbDir()` while others with no access to such `directories_` are obtained on the fly by creating such object `FileSystem::NewDirectory(..)` and manage it by unique pointers to ensure short life time. Pull Request resolved: https://github.com/facebook/rocksdb/pull/10573 Test Plan: - `make check` - Passed the repro db_stress command - For future improvement, since we currently don't assert dir containing CURRENT to be non-nullptr due to https://github.com/facebook/rocksdb/pull/10573#pullrequestreview-1087698899, there is still chances that future developers mistakenly pass down nullptr dir containing CURRENT thus resulting skipped sync dir and cause the bug again. Therefore a smarter test (e.g, such as quoted from ajkr "(make) unsynced data loss to be dropping files corresponding to unsynced directory entries") is still needed. Reviewed By: ajkr Differential Revision: D39005886 Pulled By: hx235 fbshipit-source-id: 336fb9090d0cfa6ca3dd580db86268007dde7f5a
2022-08-30 00:35:21 +00:00
}
}
if (!s.ok()) {
exec_state_ = LDBCommandExecuteResult::Failed(
"failed to unsafely remove SST file: " + s.ToString());
} else {
exec_state_ = LDBCommandExecuteResult::Succeed("unsafely removed SST file");
}
}
Add manifest fix-up utility for file temperatures (#9683) Summary: The goal of this change is to allow changes to the "current" (in FileSystem) file temperatures to feed back into DB metadata, so that they can inform decisions and stats reporting. In part because of modular code factoring, it doesn't seem easy to do this automagically, where opening an SST file and observing current Temperature different from expected would trigger a change in metadata and DB manifest write (essentially giving the deep read path access to the write path). It is also difficult to do this while the DB is open because of the limitations of LogAndApply. This change allows updating file temperature metadata on a closed DB using an experimental utility function UpdateManifestForFilesState() or `ldb update_manifest --update_temperatures`. This should suffice for "migration" scenarios where outside tooling has placed or re-arranged DB files into a (different) tiered configuration without going through RocksDB itself (currently, only compaction can change temperature metadata). Some details: * Refactored and added unit test for `ldb unsafe_remove_sst_file` because of shared functionality * Pulled in autovector.h changes from https://github.com/facebook/rocksdb/issues/9546 to fix SuperVersionContext move constructor (related to an older draft of this change) Possible follow-up work: * Support updating manifest with file checksums, such as when a new checksum function is used and want existing DB metadata updated for it. * It's possible that for some repair scenarios, lighter weight than full repair, we might want to support UpdateManifestForFilesState() to modify critical file details like size or checksum using same algorithm. But let's make sure these are differentiated from modifying file details in ways that don't suspect corruption (or require extreme trust). Pull Request resolved: https://github.com/facebook/rocksdb/pull/9683 Test Plan: unit tests added Reviewed By: jay-zhuang Differential Revision: D34798828 Pulled By: pdillinger fbshipit-source-id: cfd83e8fb10761d8c9e7f9c020d68c9106a95554
2022-03-18 23:35:51 +00:00
const std::string UpdateManifestCommand::ARG_VERBOSE = "verbose";
const std::string UpdateManifestCommand::ARG_UPDATE_TEMPERATURES =
"update_temperatures";
void UpdateManifestCommand::Help(std::string& ret) {
ret.append(" ");
ret.append(UpdateManifestCommand::Name());
ret.append(" [--update_temperatures]");
ret.append(" ");
ret.append(" MUST NOT be used on a live DB.");
ret.append("\n");
}
UpdateManifestCommand::UpdateManifestCommand(
const std::vector<std::string>& /*params*/,
const std::map<std::string, std::string>& options,
const std::vector<std::string>& flags)
: LDBCommand(options, flags, false /* is_read_only */,
BuildCmdLineOptions({ARG_VERBOSE, ARG_UPDATE_TEMPERATURES})) {
verbose_ = IsFlagPresent(flags, ARG_VERBOSE) ||
ParseBooleanOption(options, ARG_VERBOSE, false);
update_temperatures_ =
IsFlagPresent(flags, ARG_UPDATE_TEMPERATURES) ||
ParseBooleanOption(options, ARG_UPDATE_TEMPERATURES, false);
if (!update_temperatures_) {
exec_state_ = LDBCommandExecuteResult::Failed(
"No action like --update_temperatures specified for update_manifest");
}
}
void UpdateManifestCommand::DoCommand() {
PrepareOptions();
auto level = verbose_ ? InfoLogLevel::INFO_LEVEL : InfoLogLevel::WARN_LEVEL;
options_.info_log.reset(new StderrLogger(level));
experimental::UpdateManifestForFilesStateOptions opts;
opts.update_temperatures = update_temperatures_;
if (column_families_.empty()) {
column_families_.emplace_back(kDefaultColumnFamilyName, options_);
}
Status s = experimental::UpdateManifestForFilesState(options_, db_path_,
column_families_);
if (!s.ok()) {
exec_state_ = LDBCommandExecuteResult::Failed(
"failed to update manifest: " + s.ToString());
} else {
exec_state_ =
LDBCommandExecuteResult::Succeed("Manifest updates successful");
}
}
} // namespace ROCKSDB_NAMESPACE