You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
rocksdb/table/sst_file_writer.cc

344 lines
12 KiB

// 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/sst_file_writer.h"
#include <vector>
#include "db/dbformat.h"
#include "file/writable_file_writer.h"
#include "rocksdb/file_system.h"
#include "rocksdb/table.h"
#include "table/block_based/block_based_table_builder.h"
#include "table/sst_file_writer_collectors.h"
#include "test_util/sync_point.h"
namespace ROCKSDB_NAMESPACE {
const std::string ExternalSstFilePropertyNames::kVersion =
"rocksdb.external_sst_file.version";
const std::string ExternalSstFilePropertyNames::kGlobalSeqno =
"rocksdb.external_sst_file.global_seqno";
#ifndef ROCKSDB_LITE
const size_t kFadviseTrigger = 1024 * 1024; // 1MB
struct SstFileWriter::Rep {
Rep(const EnvOptions& _env_options, const Options& options,
Env::IOPriority _io_priority, const Comparator* _user_comparator,
ColumnFamilyHandle* _cfh, bool _invalidate_page_cache, bool _skip_filters)
: env_options(_env_options),
ioptions(options),
mutable_cf_options(options),
io_priority(_io_priority),
internal_comparator(_user_comparator),
cfh(_cfh),
invalidate_page_cache(_invalidate_page_cache),
last_fadvise_size(0),
skip_filters(_skip_filters) {}
std::unique_ptr<WritableFileWriter> file_writer;
std::unique_ptr<TableBuilder> builder;
EnvOptions env_options;
ImmutableOptions ioptions;
MutableCFOptions mutable_cf_options;
Env::IOPriority io_priority;
InternalKeyComparator internal_comparator;
ExternalSstFileInfo file_info;
Miscellaneous performance improvements Summary: I was investigating performance issues in the SstFileWriter and found all of the following: - The SstFileWriter::Add() function created a local InternalKey every time it was called generating a allocation and free each time. Changed to have an InternalKey member variable that can be reset with the new InternalKey::Set() function. - In SstFileWriter::Add() the smallest_key and largest_key values were assigned the result of a ToString() call, but it is simpler to just assign them directly from the user's key. - The Slice class had no move constructor so each time one was returned from a function a new one had to be allocated, the old data copied to the new, and the old one was freed. I added the move constructor which also required a copy constructor and assignment operator. - The BlockBuilder::CurrentSizeEstimate() function calculates the current estimate size, but was being called 2 or 3 times for each key added. I changed the class to maintain a running estimate (equal to the original calculation) so that the function can return an already calculated value. - The code in BlockBuilder::Add() that calculated the shared bytes between the last key and the new key duplicated what Slice::difference_offset does, so I replaced it with the standard function. - BlockBuilder::Add() had code to copy just the changed portion into the last key value (and asserted that it now matched the new key). It is more efficient just to copy the whole new key over. - Moved this same code up into the 'if (use_delta_encoding_)' since the last key value is only needed when delta encoding is on. - FlushBlockBySizePolicy::BlockAlmostFull calculated a standard deviation value each time it was called, but this information would only change if block_size of block_size_deviation changed, so I created a member variable to hold the value to avoid the calculation each time. - Each PutVarint??() function has a buffer and calls std::string::append(). Two or three calls in a row could share a buffer and a single call to std::string::append(). Some of these will be helpful outside of the SstFileWriter. I'm not 100% the addition of the move constructor is appropriate as I wonder why this wasn't done before - maybe because of compiler compatibility? I tried it on gcc 4.8 and 4.9. Test Plan: The changes should not affect the results so the existing tests should all still work and no new tests were added. The value of the changes was seen by manually testing the SstFileWriter class through MyRocks and adding timing code to identify problem areas. Reviewers: sdong, IslamAbdelRahman Reviewed By: IslamAbdelRahman Subscribers: andrewkr, dhruba Differential Revision: https://reviews.facebook.net/D59607
9 years ago
InternalKey ikey;
std::string column_family_name;
ColumnFamilyHandle* cfh;
// If true, We will give the OS a hint that this file pages is not needed
// every time we write 1MB to the file.
bool invalidate_page_cache;
// The size of the file during the last time we called Fadvise to remove
// cached pages from page cache.
uint64_t last_fadvise_size;
bool skip_filters;
Status Add(const Slice& user_key, const Slice& value,
const ValueType value_type) {
if (!builder) {
return Status::InvalidArgument("File is not opened");
}
if (file_info.num_entries == 0) {
file_info.smallest_key.assign(user_key.data(), user_key.size());
} else {
if (internal_comparator.user_comparator()->Compare(
user_key, file_info.largest_key) <= 0) {
// Make sure that keys are added in order
return Status::InvalidArgument(
"Keys must be added in strict ascending order.");
}
}
// TODO(tec) : For external SST files we could omit the seqno and type.
switch (value_type) {
case ValueType::kTypeValue:
ikey.Set(user_key, 0 /* Sequence Number */,
ValueType::kTypeValue /* Put */);
break;
case ValueType::kTypeMerge:
ikey.Set(user_key, 0 /* Sequence Number */,
ValueType::kTypeMerge /* Merge */);
break;
case ValueType::kTypeDeletion:
ikey.Set(user_key, 0 /* Sequence Number */,
ValueType::kTypeDeletion /* Delete */);
break;
default:
return Status::InvalidArgument("Value type is not supported");
}
builder->Add(ikey.Encode(), value);
// update file info
file_info.num_entries++;
file_info.largest_key.assign(user_key.data(), user_key.size());
file_info.file_size = builder->FileSize();
InvalidatePageCache(false /* closing */).PermitUncheckedError();
return Status::OK();
}
Status DeleteRange(const Slice& begin_key, const Slice& end_key) {
if (!builder) {
return Status::InvalidArgument("File is not opened");
}
RangeTombstone tombstone(begin_key, end_key, 0 /* Sequence Number */);
if (file_info.num_range_del_entries == 0) {
file_info.smallest_range_del_key.assign(tombstone.start_key_.data(),
tombstone.start_key_.size());
file_info.largest_range_del_key.assign(tombstone.end_key_.data(),
tombstone.end_key_.size());
} else {
if (internal_comparator.user_comparator()->Compare(
tombstone.start_key_, file_info.smallest_range_del_key) < 0) {
file_info.smallest_range_del_key.assign(tombstone.start_key_.data(),
tombstone.start_key_.size());
}
if (internal_comparator.user_comparator()->Compare(
tombstone.end_key_, file_info.largest_range_del_key) > 0) {
file_info.largest_range_del_key.assign(tombstone.end_key_.data(),
tombstone.end_key_.size());
}
}
auto ikey_and_end_key = tombstone.Serialize();
builder->Add(ikey_and_end_key.first.Encode(), ikey_and_end_key.second);
// update file info
file_info.num_range_del_entries++;
file_info.file_size = builder->FileSize();
InvalidatePageCache(false /* closing */).PermitUncheckedError();
return Status::OK();
}
Status InvalidatePageCache(bool closing) {
Status s = Status::OK();
if (invalidate_page_cache == false) {
// Fadvise disabled
return s;
}
uint64_t bytes_since_last_fadvise =
builder->FileSize() - last_fadvise_size;
if (bytes_since_last_fadvise > kFadviseTrigger || closing) {
TEST_SYNC_POINT_CALLBACK("SstFileWriter::Rep::InvalidatePageCache",
&(bytes_since_last_fadvise));
// Tell the OS that we don't need this file in page cache
s = file_writer->InvalidateCache(0, 0);
if (s.IsNotSupported()) {
// NotSupported is fine as it could be a file type that doesn't use page
// cache.
s = Status::OK();
}
last_fadvise_size = builder->FileSize();
}
return s;
}
};
SstFileWriter::SstFileWriter(const EnvOptions& env_options,
const Options& options,
const Comparator* user_comparator,
ColumnFamilyHandle* column_family,
bool invalidate_page_cache,
Env::IOPriority io_priority, bool skip_filters)
: rep_(new Rep(env_options, options, io_priority, user_comparator,
column_family, invalidate_page_cache, skip_filters)) {
rep_->file_info.file_size = 0;
}
SstFileWriter::~SstFileWriter() {
if (rep_->builder) {
// User did not call Finish() or Finish() failed, we need to
// abandon the builder.
rep_->builder->Abandon();
}
}
Status SstFileWriter::Open(const std::string& file_path) {
Rep* r = rep_.get();
Status s;
std::unique_ptr<FSWritableFile> sst_file;
FileOptions cur_file_opts(r->env_options);
s = r->ioptions.env->GetFileSystem()->NewWritableFile(
file_path, cur_file_opts, &sst_file, nullptr);
if (!s.ok()) {
return s;
}
sst_file->SetIOPriority(r->io_priority);
CompressionType compression_type;
CompressionOptions compression_opts;
if (r->mutable_cf_options.bottommost_compression !=
kDisableCompressionOption) {
compression_type = r->mutable_cf_options.bottommost_compression;
if (r->mutable_cf_options.bottommost_compression_opts.enabled) {
compression_opts = r->mutable_cf_options.bottommost_compression_opts;
} else {
compression_opts = r->mutable_cf_options.compression_opts;
}
} else if (!r->ioptions.compression_per_level.empty()) {
// Use the compression of the last level if we have per level compression
compression_type = *(r->ioptions.compression_per_level.rbegin());
compression_opts = r->mutable_cf_options.compression_opts;
} else {
compression_type = r->mutable_cf_options.compression;
compression_opts = r->mutable_cf_options.compression_opts;
}
IntTblPropCollectorFactories int_tbl_prop_collector_factories;
// SstFileWriter properties collector to add SstFileWriter version.
int_tbl_prop_collector_factories.emplace_back(
new SstFileWriterPropertiesCollectorFactory(2 /* version */,
0 /* global_seqno*/));
// User collector factories
auto user_collector_factories =
r->ioptions.table_properties_collector_factories;
for (size_t i = 0; i < user_collector_factories.size(); i++) {
int_tbl_prop_collector_factories.emplace_back(
new UserKeyTablePropertiesCollectorFactory(
user_collector_factories[i]));
}
int unknown_level = -1;
uint32_t cf_id;
if (r->cfh != nullptr) {
// user explicitly specified that this file will be ingested into cfh,
// we can persist this information in the file.
cf_id = r->cfh->GetID();
r->column_family_name = r->cfh->GetName();
} else {
r->column_family_name = "";
cf_id = TablePropertiesCollectorFactory::Context::kUnknownColumnFamily;
}
// SstFileWriter is used to create sst files that can be added to database
// later. Therefore, no real db_id and db_session_id are associated with it.
// Here we mimic the way db_session_id behaves by resetting the db_session_id
// every time SstFileWriter is used, and in this case db_id is set to be "SST
// Writer".
std::string db_session_id = r->ioptions.env->GenerateUniqueId();
if (!db_session_id.empty() && db_session_id.back() == '\n') {
db_session_id.pop_back();
}
TableBuilderOptions table_builder_options(
r->ioptions, r->mutable_cf_options, r->internal_comparator,
&int_tbl_prop_collector_factories, compression_type, compression_opts,
Add more LSM info to FilterBuildingContext (#8246) Summary: Add `num_levels`, `is_bottommost`, and table file creation `reason` to `FilterBuildingContext`, in anticipation of more powerful Bloom-like filter support. To support this, added `is_bottommost` and `reason` to `TableBuilderOptions`, which allowed removing `reason` parameter from `rocksdb::BuildTable`. I attempted to remove `skip_filters` from `TableBuilderOptions`, because filter construction decisions should arise from options, not one-off parameters. I could not completely remove it because the public API for SstFileWriter takes a `skip_filters` parameter, and translating this into an option change would mean awkwardly replacing the table_factory if it is BlockBasedTableFactory with new filter_policy=nullptr option. I marked this public skip_filters option as deprecated because of this oddity. (skip_filters on the read side probably makes sense.) At least `skip_filters` is now largely hidden for users of `TableBuilderOptions` and is no longer used for implementing the optimize_filters_for_hits option. Bringing the logic for that option closer to handling of FilterBuildingContext makes it more obvious that hese two are using the same notion of "bottommost." (Planned: configuration options for Bloom-like filters that generalize `optimize_filters_for_hits`) Recommended follow-up: Try to get away from "bottommost level" naming of things, which is inaccurate (see VersionStorageInfo::RangeMightExistAfterSortedRun), and move to "bottommost run" or just "bottommost." Pull Request resolved: https://github.com/facebook/rocksdb/pull/8246 Test Plan: extended an existing unit test to exercise and check various filter building contexts. Also, existing tests for optimize_filters_for_hits validate some of the "bottommost" handling, which is now closely connected to FilterBuildingContext::is_bottommost through TableBuilderOptions::is_bottommost Reviewed By: mrambacher Differential Revision: D28099346 Pulled By: pdillinger fbshipit-source-id: 2c1072e29c24d4ac404c761a7b7663292372600a
4 years ago
cf_id, r->column_family_name, unknown_level, false /* is_bottommost */,
TableFileCreationReason::kMisc, 0 /* creation_time */,
0 /* oldest_key_time */, 0 /* file_creation_time */,
"SST Writer" /* db_id */, db_session_id, 0 /* target_file_size */, 0);
Add more LSM info to FilterBuildingContext (#8246) Summary: Add `num_levels`, `is_bottommost`, and table file creation `reason` to `FilterBuildingContext`, in anticipation of more powerful Bloom-like filter support. To support this, added `is_bottommost` and `reason` to `TableBuilderOptions`, which allowed removing `reason` parameter from `rocksdb::BuildTable`. I attempted to remove `skip_filters` from `TableBuilderOptions`, because filter construction decisions should arise from options, not one-off parameters. I could not completely remove it because the public API for SstFileWriter takes a `skip_filters` parameter, and translating this into an option change would mean awkwardly replacing the table_factory if it is BlockBasedTableFactory with new filter_policy=nullptr option. I marked this public skip_filters option as deprecated because of this oddity. (skip_filters on the read side probably makes sense.) At least `skip_filters` is now largely hidden for users of `TableBuilderOptions` and is no longer used for implementing the optimize_filters_for_hits option. Bringing the logic for that option closer to handling of FilterBuildingContext makes it more obvious that hese two are using the same notion of "bottommost." (Planned: configuration options for Bloom-like filters that generalize `optimize_filters_for_hits`) Recommended follow-up: Try to get away from "bottommost level" naming of things, which is inaccurate (see VersionStorageInfo::RangeMightExistAfterSortedRun), and move to "bottommost run" or just "bottommost." Pull Request resolved: https://github.com/facebook/rocksdb/pull/8246 Test Plan: extended an existing unit test to exercise and check various filter building contexts. Also, existing tests for optimize_filters_for_hits validate some of the "bottommost" handling, which is now closely connected to FilterBuildingContext::is_bottommost through TableBuilderOptions::is_bottommost Reviewed By: mrambacher Differential Revision: D28099346 Pulled By: pdillinger fbshipit-source-id: 2c1072e29c24d4ac404c761a7b7663292372600a
4 years ago
// XXX: when we can remove skip_filters from the SstFileWriter public API
// we can remove it from TableBuilderOptions.
table_builder_options.skip_filters = r->skip_filters;
FileTypeSet tmp_set = r->ioptions.checksum_handoff_file_types;
r->file_writer.reset(new WritableFileWriter(
std::move(sst_file), file_path, r->env_options, r->ioptions.clock,
nullptr /* io_tracer */, nullptr /* stats */, r->ioptions.listeners,
r->ioptions.file_checksum_gen_factory.get(),
Using existing crc32c checksum in checksum handoff for Manifest and WAL (#8412) Summary: In PR https://github.com/facebook/rocksdb/issues/7523 , checksum handoff is introduced in RocksDB for WAL, Manifest, and SST files. When user enable checksum handoff for a certain type of file, before the data is written to the lower layer storage system, we calculate the checksum (crc32c) of each piece of data and pass the checksum down with the data, such that data verification can be down by the lower layer storage system if it has the capability. However, it cannot cover the whole lifetime of the data in the memory and also it potentially introduces extra checksum calculation overhead. In this PR, we introduce a new interface in WritableFileWriter::Append, which allows the caller be able to pass the data and the checksum (crc32c) together. In this way, WritableFileWriter can directly use the pass-in checksum (crc32c) to generate the checksum of data being passed down to the storage system. It saves the calculation overhead and achieves higher protection coverage. When a new checksum is added with the data, we use Crc32cCombine https://github.com/facebook/rocksdb/issues/8305 to combine the existing checksum and the new checksum. To avoid the segmenting of data by rate-limiter before it is stored, rate-limiter is called enough times to accumulate enough credits for a certain write. This design only support Manifest and WAL which use log_writer in the current stage. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8412 Test Plan: make check, add new testing cases. Reviewed By: anand1976 Differential Revision: D29151545 Pulled By: zhichao-cao fbshipit-source-id: 75e2278c5126cfd58393c67b1efd18dcc7a30772
4 years ago
tmp_set.Contains(FileType::kTableFile), false));
// TODO(tec) : If table_factory is using compressed block cache, we will
// be adding the external sst file blocks into it, which is wasteful.
r->builder.reset(r->ioptions.table_factory->NewTableBuilder(
table_builder_options, r->file_writer.get()));
r->file_info = ExternalSstFileInfo();
r->file_info.file_path = file_path;
r->file_info.version = 2;
return s;
}
Status SstFileWriter::Add(const Slice& user_key, const Slice& value) {
return rep_->Add(user_key, value, ValueType::kTypeValue);
}
Status SstFileWriter::Put(const Slice& user_key, const Slice& value) {
return rep_->Add(user_key, value, ValueType::kTypeValue);
}
Status SstFileWriter::Merge(const Slice& user_key, const Slice& value) {
return rep_->Add(user_key, value, ValueType::kTypeMerge);
}
Status SstFileWriter::Delete(const Slice& user_key) {
return rep_->Add(user_key, Slice(), ValueType::kTypeDeletion);
}
Status SstFileWriter::DeleteRange(const Slice& begin_key,
const Slice& end_key) {
return rep_->DeleteRange(begin_key, end_key);
}
Status SstFileWriter::Finish(ExternalSstFileInfo* file_info) {
Rep* r = rep_.get();
if (!r->builder) {
return Status::InvalidArgument("File is not opened");
}
if (r->file_info.num_entries == 0 &&
r->file_info.num_range_del_entries == 0) {
return Status::InvalidArgument("Cannot create sst file with no entries");
}
Status s = r->builder->Finish();
r->file_info.file_size = r->builder->FileSize();
if (s.ok()) {
s = r->file_writer->Sync(r->ioptions.use_fsync);
r->InvalidatePageCache(true /* closing */).PermitUncheckedError();
if (s.ok()) {
s = r->file_writer->Close();
}
}
if (s.ok()) {
r->file_info.file_checksum = r->file_writer->GetFileChecksum();
r->file_info.file_checksum_func_name =
r->file_writer->GetFileChecksumFuncName();
}
if (!s.ok()) {
r->ioptions.env->DeleteFile(r->file_info.file_path);
}
if (file_info != nullptr) {
*file_info = r->file_info;
}
r->builder.reset();
return s;
}
uint64_t SstFileWriter::FileSize() {
return rep_->file_info.file_size;
}
#endif // !ROCKSDB_LITE
} // namespace ROCKSDB_NAMESPACE