rocksdb/db/version_set_sync_and_async.h
Changyu Bi 62fc15f009 Block per key-value checksum (#11287)
Summary:
add option `block_protection_bytes_per_key` and implementation for block per key-value checksum. The main changes are
1. checksum construction and verification in block.cc/h
2. pass the option `block_protection_bytes_per_key` around (mainly for methods defined in table_cache.h)
3. unit tests/crash test updates

Tests:
* Added unit tests
* Crash test: `python3 tools/db_crashtest.py blackbox --simple --block_protection_bytes_per_key=1 --write_buffer_size=1048576`

Follow up (maybe as a separate PR): make sure corruption status returned from BlockIters are correctly handled.

Performance:
Turning on block per KV protection has a non-trivial negative impact on read performance and costs additional memory.
For memory, each block includes additional 24 bytes for checksum-related states beside checksum itself. For CPU, I set up a DB of size ~1.2GB with 5M keys (32 bytes key and 200 bytes value) which compacts to ~5 SST files (target file size 256 MB) in L6 without compression. I tested readrandom performance with various block cache size (to mimic various cache hit rates):

```
SETUP
make OPTIMIZE_LEVEL="-O3" USE_LTO=1 DEBUG_LEVEL=0 -j32 db_bench
./db_bench -benchmarks=fillseq,compact0,waitforcompaction,compact,waitforcompaction -write_buffer_size=33554432 -level_compaction_dynamic_level_bytes=true -max_background_jobs=8 -target_file_size_base=268435456 --num=5000000 --key_size=32 --value_size=200 --compression_type=none

BENCHMARK
./db_bench --use_existing_db -benchmarks=readtocache,readrandom[-X10] --num=5000000 --key_size=32 --disable_auto_compactions --reads=1000000 --block_protection_bytes_per_key=[0|1] --cache_size=$CACHESIZE

The readrandom ops/sec looks like the following:
Block cache size:  2GB        1.2GB * 0.9    1.2GB * 0.8     1.2GB * 0.5   8MB
Main              240805     223604         198176           161653       139040
PR prot_bytes=0   238691     226693         200127           161082       141153
PR prot_bytes=1   214983     193199         178532           137013       108211
prot_bytes=1 vs    -10%        -15%          -10.8%          -15%        -23%
prot_bytes=0
```

The benchmark has a lot of variance, but there was a 5% to 25% regression in this benchmark with different cache hit rates.

Pull Request resolved: https://github.com/facebook/rocksdb/pull/11287

Reviewed By: ajkr

Differential Revision: D43970708

Pulled By: cbi42

fbshipit-source-id: ef98d898b71779846fa74212b9ec9e08b7183940
2023-04-25 12:08:23 -07:00

173 lines
6.2 KiB
C++

// Copyright (c) Meta Platforms, Inc. and affiliates.
//
// 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 "util/coro_utils.h"
#if defined(WITHOUT_COROUTINES) || \
(defined(USE_COROUTINES) && defined(WITH_COROUTINES))
namespace ROCKSDB_NAMESPACE {
// Lookup a batch of keys in a single SST file
DEFINE_SYNC_AND_ASYNC(Status, Version::MultiGetFromSST)
(const ReadOptions& read_options, MultiGetRange file_range, int hit_file_level,
bool skip_filters, bool skip_range_deletions, FdWithKeyRange* f,
std::unordered_map<uint64_t, BlobReadContexts>& blob_ctxs,
TableCache::TypedHandle* table_handle, uint64_t& num_filter_read,
uint64_t& num_index_read, uint64_t& num_sst_read) {
bool timer_enabled = GetPerfLevel() >= PerfLevel::kEnableTimeExceptForMutex &&
get_perf_context()->per_level_perf_context_enabled;
Status s;
StopWatchNano timer(clock_, timer_enabled /* auto_start */);
s = CO_AWAIT(table_cache_->MultiGet)(
read_options, *internal_comparator(), *f->file_metadata, &file_range,
mutable_cf_options_.block_protection_bytes_per_key,
mutable_cf_options_.prefix_extractor,
cfd_->internal_stats()->GetFileReadHist(hit_file_level), skip_filters,
skip_range_deletions, hit_file_level, table_handle);
// TODO: examine the behavior for corrupted key
if (timer_enabled) {
PERF_COUNTER_BY_LEVEL_ADD(get_from_table_nanos, timer.ElapsedNanos(),
hit_file_level);
}
if (!s.ok()) {
// TODO: Set status for individual keys appropriately
for (auto iter = file_range.begin(); iter != file_range.end(); ++iter) {
*iter->s = s;
file_range.MarkKeyDone(iter);
}
CO_RETURN s;
}
uint64_t batch_size = 0;
for (auto iter = file_range.begin(); s.ok() && iter != file_range.end();
++iter) {
GetContext& get_context = *iter->get_context;
Status* status = iter->s;
// The Status in the KeyContext takes precedence over GetContext state
// Status may be an error if there were any IO errors in the table
// reader. We never expect Status to be NotFound(), as that is
// determined by get_context
assert(!status->IsNotFound());
if (!status->ok()) {
file_range.MarkKeyDone(iter);
continue;
}
if (get_context.sample()) {
sample_file_read_inc(f->file_metadata);
}
batch_size++;
num_index_read += get_context.get_context_stats_.num_index_read;
num_filter_read += get_context.get_context_stats_.num_filter_read;
num_sst_read += get_context.get_context_stats_.num_sst_read;
// Reset these stats since they're specific to a level
get_context.get_context_stats_.num_index_read = 0;
get_context.get_context_stats_.num_filter_read = 0;
get_context.get_context_stats_.num_sst_read = 0;
// report the counters before returning
if (get_context.State() != GetContext::kNotFound &&
get_context.State() != GetContext::kMerge &&
db_statistics_ != nullptr) {
get_context.ReportCounters();
} else {
if (iter->max_covering_tombstone_seq > 0) {
// The remaining files we look at will only contain covered keys, so
// we stop here for this key
file_range.SkipKey(iter);
}
}
switch (get_context.State()) {
case GetContext::kNotFound:
// Keep searching in other files
break;
case GetContext::kMerge:
// TODO: update per-level perfcontext user_key_return_count for kMerge
break;
case GetContext::kFound:
if (hit_file_level == 0) {
RecordTick(db_statistics_, GET_HIT_L0);
} else if (hit_file_level == 1) {
RecordTick(db_statistics_, GET_HIT_L1);
} else if (hit_file_level >= 2) {
RecordTick(db_statistics_, GET_HIT_L2_AND_UP);
}
PERF_COUNTER_BY_LEVEL_ADD(user_key_return_count, 1, hit_file_level);
file_range.MarkKeyDone(iter);
if (iter->is_blob_index) {
BlobIndex blob_index;
Status tmp_s;
if (iter->value) {
TEST_SYNC_POINT_CALLBACK("Version::MultiGet::TamperWithBlobIndex",
&(*iter));
tmp_s = blob_index.DecodeFrom(*(iter->value));
} else {
assert(iter->columns);
assert(!iter->columns->columns().empty());
assert(iter->columns->columns().front().name() ==
kDefaultWideColumnName);
tmp_s =
blob_index.DecodeFrom(iter->columns->columns().front().value());
}
if (tmp_s.ok()) {
const uint64_t blob_file_num = blob_index.file_number();
blob_ctxs[blob_file_num].emplace_back(blob_index, &*iter);
} else {
*(iter->s) = tmp_s;
}
} else {
if (iter->value) {
file_range.AddValueSize(iter->value->size());
} else {
assert(iter->columns);
file_range.AddValueSize(iter->columns->serialized_size());
}
if (file_range.GetValueSize() > read_options.value_size_soft_limit) {
s = Status::Aborted();
break;
}
}
continue;
case GetContext::kDeleted:
// Use empty error message for speed
*status = Status::NotFound();
file_range.MarkKeyDone(iter);
continue;
case GetContext::kCorrupt:
*status =
Status::Corruption("corrupted key for ", iter->lkey->user_key());
file_range.MarkKeyDone(iter);
continue;
case GetContext::kUnexpectedBlobIndex:
ROCKS_LOG_ERROR(info_log_, "Encounter unexpected blob index.");
*status = Status::NotSupported(
"Encounter unexpected blob index. Please open DB with "
"ROCKSDB_NAMESPACE::blob_db::BlobDB instead.");
file_range.MarkKeyDone(iter);
continue;
case GetContext::kMergeOperatorFailed:
*status = Status::Corruption(Status::SubCode::kMergeOperatorFailed);
file_range.MarkKeyDone(iter);
continue;
}
}
RecordInHistogram(db_statistics_, SST_BATCH_SIZE, batch_size);
CO_RETURN s;
}
} // namespace ROCKSDB_NAMESPACE
#endif