2016-02-09 23:12:00 +00:00
|
|
|
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
2017-07-15 23:03:42 +00:00
|
|
|
// 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).
|
2014-11-14 19:38:26 +00:00
|
|
|
//
|
2019-05-30 03:44:08 +00:00
|
|
|
#include "file/file_util.h"
|
2015-08-19 22:02:17 +00:00
|
|
|
|
2014-11-14 19:38:26 +00:00
|
|
|
#include <string>
|
|
|
|
#include <algorithm>
|
2015-08-19 22:02:17 +00:00
|
|
|
|
2019-09-16 17:31:27 +00:00
|
|
|
#include "file/random_access_file_reader.h"
|
|
|
|
#include "file/sequence_file_reader.h"
|
2019-05-30 03:44:08 +00:00
|
|
|
#include "file/sst_file_manager_impl.h"
|
2019-09-16 17:31:27 +00:00
|
|
|
#include "file/writable_file_writer.h"
|
2014-11-14 19:38:26 +00:00
|
|
|
#include "rocksdb/env.h"
|
|
|
|
|
|
|
|
namespace rocksdb {
|
|
|
|
|
|
|
|
// Utility function to copy a file up to a specified length
|
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
|
|
|
Status CopyFile(FileSystem* fs, const std::string& source,
|
2016-12-29 02:38:20 +00:00
|
|
|
const std::string& destination, uint64_t size, bool use_fsync) {
|
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
|
|
|
const FileOptions soptions;
|
2014-11-14 19:38:26 +00:00
|
|
|
Status s;
|
2018-11-09 19:17:34 +00:00
|
|
|
std::unique_ptr<SequentialFileReader> src_reader;
|
|
|
|
std::unique_ptr<WritableFileWriter> dest_writer;
|
Move rate_limiter, write buffering, most perf context instrumentation and most random kill out of Env
Summary: We want to keep Env a think layer for better portability. Less platform dependent codes should be moved out of Env. In this patch, I create a wrapper of file readers and writers, and put rate limiting, write buffering, as well as most perf context instrumentation and random kill out of Env. It will make it easier to maintain multiple Env in the future.
Test Plan: Run all existing unit tests.
Reviewers: anthony, kradhakrishnan, IslamAbdelRahman, yhchiang, igor
Reviewed By: igor
Subscribers: leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D42321
2015-07-17 23:16:11 +00:00
|
|
|
|
|
|
|
{
|
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
|
|
|
std::unique_ptr<FSSequentialFile> srcfile;
|
|
|
|
s = fs->NewSequentialFile(source, soptions, &srcfile, nullptr);
|
2017-11-03 18:48:12 +00:00
|
|
|
if (!s.ok()) {
|
|
|
|
return s;
|
|
|
|
}
|
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
|
|
|
std::unique_ptr<FSWritableFile> destfile;
|
|
|
|
s = fs->NewWritableFile(destination, soptions, &destfile, nullptr);
|
2017-11-03 18:48:12 +00:00
|
|
|
if (!s.ok()) {
|
|
|
|
return s;
|
|
|
|
}
|
2014-11-14 19:38:26 +00:00
|
|
|
|
2017-11-03 18:48:12 +00:00
|
|
|
if (size == 0) {
|
|
|
|
// default argument means copy everything
|
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
|
|
|
s = fs->GetFileSize(source, IOOptions(), &size, nullptr);
|
2017-11-03 18:48:12 +00:00
|
|
|
if (!s.ok()) {
|
|
|
|
return s;
|
|
|
|
}
|
2014-11-14 19:38:26 +00:00
|
|
|
}
|
2018-06-21 15:34:24 +00:00
|
|
|
src_reader.reset(new SequentialFileReader(std::move(srcfile), source));
|
2018-08-23 17:04:10 +00:00
|
|
|
dest_writer.reset(
|
|
|
|
new WritableFileWriter(std::move(destfile), destination, soptions));
|
Move rate_limiter, write buffering, most perf context instrumentation and most random kill out of Env
Summary: We want to keep Env a think layer for better portability. Less platform dependent codes should be moved out of Env. In this patch, I create a wrapper of file readers and writers, and put rate limiting, write buffering, as well as most perf context instrumentation and random kill out of Env. It will make it easier to maintain multiple Env in the future.
Test Plan: Run all existing unit tests.
Reviewers: anthony, kradhakrishnan, IslamAbdelRahman, yhchiang, igor
Reviewed By: igor
Subscribers: leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D42321
2015-07-17 23:16:11 +00:00
|
|
|
}
|
2014-11-14 19:38:26 +00:00
|
|
|
|
|
|
|
char buffer[4096];
|
|
|
|
Slice slice;
|
|
|
|
while (size > 0) {
|
2015-10-19 20:40:44 +00:00
|
|
|
size_t bytes_to_read = std::min(sizeof(buffer), static_cast<size_t>(size));
|
2017-11-03 18:48:12 +00:00
|
|
|
s = src_reader->Read(bytes_to_read, &slice, buffer);
|
|
|
|
if (!s.ok()) {
|
|
|
|
return s;
|
2014-11-14 19:38:26 +00:00
|
|
|
}
|
2017-11-03 18:48:12 +00:00
|
|
|
if (slice.size() == 0) {
|
|
|
|
return Status::Corruption("file too small");
|
2014-11-14 19:38:26 +00:00
|
|
|
}
|
2017-11-03 18:48:12 +00:00
|
|
|
s = dest_writer->Append(slice);
|
2014-11-14 19:38:26 +00:00
|
|
|
if (!s.ok()) {
|
|
|
|
return s;
|
|
|
|
}
|
|
|
|
size -= slice.size();
|
|
|
|
}
|
2018-04-19 21:00:52 +00:00
|
|
|
return dest_writer->Sync(use_fsync);
|
2014-11-14 19:38:26 +00:00
|
|
|
}
|
|
|
|
|
2016-03-17 17:07:21 +00:00
|
|
|
// Utility function to create a file with the provided contents
|
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
|
|
|
Status CreateFile(FileSystem* fs, const std::string& destination,
|
2018-08-28 19:35:17 +00:00
|
|
|
const std::string& contents, bool use_fsync) {
|
2016-03-17 17:07:21 +00:00
|
|
|
const EnvOptions soptions;
|
|
|
|
Status s;
|
2018-11-09 19:17:34 +00:00
|
|
|
std::unique_ptr<WritableFileWriter> dest_writer;
|
2016-03-17 17:07:21 +00:00
|
|
|
|
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
|
|
|
std::unique_ptr<FSWritableFile> destfile;
|
|
|
|
s = fs->NewWritableFile(destination, soptions, &destfile, nullptr);
|
2016-03-17 17:07:21 +00:00
|
|
|
if (!s.ok()) {
|
|
|
|
return s;
|
|
|
|
}
|
2018-08-23 17:04:10 +00:00
|
|
|
dest_writer.reset(
|
|
|
|
new WritableFileWriter(std::move(destfile), destination, soptions));
|
2018-08-28 19:35:17 +00:00
|
|
|
s = dest_writer->Append(Slice(contents));
|
|
|
|
if (!s.ok()) {
|
|
|
|
return s;
|
|
|
|
}
|
|
|
|
return dest_writer->Sync(use_fsync);
|
2016-03-17 17:07:21 +00:00
|
|
|
}
|
|
|
|
|
2019-01-29 22:27:30 +00:00
|
|
|
Status DeleteDBFile(const ImmutableDBOptions* db_options,
|
2019-07-07 04:04:22 +00:00
|
|
|
const std::string& fname, const std::string& dir_to_sync,
|
|
|
|
const bool force_bg, const bool force_fg) {
|
2016-12-22 01:35:00 +00:00
|
|
|
#ifndef ROCKSDB_LITE
|
2019-03-28 22:13:02 +00:00
|
|
|
SstFileManagerImpl* sfm =
|
2016-01-29 02:35:01 +00:00
|
|
|
static_cast<SstFileManagerImpl*>(db_options->sst_file_manager.get());
|
2019-07-07 04:04:22 +00:00
|
|
|
if (sfm && !force_fg) {
|
2019-01-29 22:27:30 +00:00
|
|
|
return sfm->ScheduleFileDeletion(fname, dir_to_sync, force_bg);
|
2015-08-19 22:02:17 +00:00
|
|
|
} else {
|
2016-01-29 02:35:01 +00:00
|
|
|
return db_options->env->DeleteFile(fname);
|
2015-08-19 22:02:17 +00:00
|
|
|
}
|
2016-12-22 01:35:00 +00:00
|
|
|
#else
|
2018-04-26 20:51:39 +00:00
|
|
|
(void)dir_to_sync;
|
2019-01-29 22:27:30 +00:00
|
|
|
(void)force_bg;
|
2019-07-07 04:04:22 +00:00
|
|
|
(void)force_fg;
|
2016-12-22 01:35:00 +00:00
|
|
|
// SstFileManager is not supported in ROCKSDB_LITE
|
2019-03-28 22:13:02 +00:00
|
|
|
// Delete file immediately
|
2016-12-22 01:35:00 +00:00
|
|
|
return db_options->env->DeleteFile(fname);
|
|
|
|
#endif
|
2015-08-19 22:02:17 +00:00
|
|
|
}
|
|
|
|
|
2019-07-07 04:04:22 +00:00
|
|
|
bool IsWalDirSameAsDBPath(const ImmutableDBOptions* db_options) {
|
|
|
|
bool same = false;
|
2019-07-09 18:01:12 +00:00
|
|
|
assert(!db_options->db_paths.empty());
|
2019-07-07 04:04:22 +00:00
|
|
|
Status s = db_options->env->AreFilesSame(db_options->wal_dir,
|
|
|
|
db_options->db_paths[0].path, &same);
|
|
|
|
if (s.IsNotSupported()) {
|
|
|
|
same = db_options->wal_dir == db_options->db_paths[0].path;
|
|
|
|
}
|
|
|
|
return same;
|
|
|
|
}
|
|
|
|
|
2014-11-14 19:38:26 +00:00
|
|
|
} // namespace rocksdb
|