2016-02-10 00:12:00 +01:00
|
|
|
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
2017-07-16 01:03:42 +02:00
|
|
|
// 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).
|
2013-12-05 22:09:13 +01:00
|
|
|
#include "table/meta_blocks.h"
|
|
|
|
|
|
|
|
#include <map>
|
2014-02-05 01:21:47 +01:00
|
|
|
#include <string>
|
2013-12-05 22:09:13 +01:00
|
|
|
|
2019-05-30 23:47:29 +02:00
|
|
|
#include "block_fetcher.h"
|
A new call back to TablePropertiesCollector to allow users know the entry is add, delete or merge
Summary:
Currently users have no idea a key is add, delete or merge from TablePropertiesCollector call back. Add a new function to add it.
Also refactor the codes so that
(1) make table property collector and internal table property collector two separate data structures with the later one now exposed
(2) table builders only receive internal table properties
Test Plan: Add cases in table_properties_collector_test to cover both of old and new ways of using TablePropertiesCollector.
Reviewers: yhchiang, igor.sugak, rven, igor
Reviewed By: rven, igor
Subscribers: meyering, yoshinorim, maykov, leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D35373
2015-04-06 19:04:30 +02:00
|
|
|
#include "db/table_properties_collector.h"
|
2019-09-16 19:31:27 +02:00
|
|
|
#include "file/random_access_file_reader.h"
|
2014-01-28 06:58:46 +01:00
|
|
|
#include "rocksdb/table.h"
|
2014-02-05 01:21:47 +01:00
|
|
|
#include "rocksdb/table_properties.h"
|
2019-05-30 23:47:29 +02:00
|
|
|
#include "table/block_based/block.h"
|
2013-12-05 22:09:13 +01:00
|
|
|
#include "table/format.h"
|
2015-10-13 00:06:38 +02:00
|
|
|
#include "table/internal_iterator.h"
|
2015-12-16 03:20:10 +01:00
|
|
|
#include "table/persistent_cache_helper.h"
|
2014-11-13 20:39:30 +01:00
|
|
|
#include "table/table_properties_internal.h"
|
2019-05-31 02:39:43 +02:00
|
|
|
#include "test_util/sync_point.h"
|
2013-12-05 22:09:13 +01:00
|
|
|
#include "util/coding.h"
|
|
|
|
|
2020-02-20 21:07:53 +01:00
|
|
|
namespace ROCKSDB_NAMESPACE {
|
2013-12-05 22:09:13 +01:00
|
|
|
|
|
|
|
MetaIndexBuilder::MetaIndexBuilder()
|
2017-11-21 01:31:04 +01:00
|
|
|
: meta_index_block_(new BlockBuilder(1 /* restart interval */)) {}
|
2013-12-05 22:09:13 +01:00
|
|
|
|
|
|
|
void MetaIndexBuilder::Add(const std::string& key,
|
|
|
|
const BlockHandle& handle) {
|
|
|
|
std::string handle_encoding;
|
|
|
|
handle.EncodeTo(&handle_encoding);
|
|
|
|
meta_block_handles_.insert({key, handle_encoding});
|
|
|
|
}
|
|
|
|
|
|
|
|
Slice MetaIndexBuilder::Finish() {
|
|
|
|
for (const auto& metablock : meta_block_handles_) {
|
|
|
|
meta_index_block_->Add(metablock.first, metablock.second);
|
|
|
|
}
|
|
|
|
return meta_index_block_->Finish();
|
|
|
|
}
|
|
|
|
|
2018-06-12 21:46:48 +02:00
|
|
|
// Property block will be read sequentially and cached in a heap located
|
|
|
|
// object, so there's no need for restart points. Thus we set the restart
|
|
|
|
// interval to infinity to save space.
|
2013-12-05 22:09:13 +01:00
|
|
|
PropertyBlockBuilder::PropertyBlockBuilder()
|
2018-06-12 21:46:48 +02:00
|
|
|
: properties_block_(
|
|
|
|
new BlockBuilder(port::kMaxInt32 /* restart interval */)) {}
|
2013-12-05 22:09:13 +01:00
|
|
|
|
|
|
|
void PropertyBlockBuilder::Add(const std::string& name,
|
|
|
|
const std::string& val) {
|
|
|
|
props_.insert({name, val});
|
|
|
|
}
|
|
|
|
|
|
|
|
void PropertyBlockBuilder::Add(const std::string& name, uint64_t val) {
|
|
|
|
assert(props_.find(name) == props_.end());
|
|
|
|
|
|
|
|
std::string dst;
|
|
|
|
PutVarint64(&dst, val);
|
|
|
|
|
|
|
|
Add(name, dst);
|
|
|
|
}
|
|
|
|
|
|
|
|
void PropertyBlockBuilder::Add(
|
|
|
|
const UserCollectedProperties& user_collected_properties) {
|
|
|
|
for (const auto& prop : user_collected_properties) {
|
|
|
|
Add(prop.first, prop.second);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
void PropertyBlockBuilder::AddTableProperty(const TableProperties& props) {
|
2019-04-19 07:36:32 +02:00
|
|
|
TEST_SYNC_POINT_CALLBACK("PropertyBlockBuilder::AddTableProperty:Start",
|
|
|
|
const_cast<TableProperties*>(&props));
|
|
|
|
|
2013-12-05 22:09:13 +01:00
|
|
|
Add(TablePropertiesNames::kRawKeySize, props.raw_key_size);
|
|
|
|
Add(TablePropertiesNames::kRawValueSize, props.raw_value_size);
|
|
|
|
Add(TablePropertiesNames::kDataSize, props.data_size);
|
|
|
|
Add(TablePropertiesNames::kIndexSize, props.index_size);
|
2017-06-13 19:59:22 +02:00
|
|
|
if (props.index_partitions != 0) {
|
|
|
|
Add(TablePropertiesNames::kIndexPartitions, props.index_partitions);
|
|
|
|
Add(TablePropertiesNames::kTopLevelIndexSize, props.top_level_index_size);
|
|
|
|
}
|
2018-05-26 03:41:31 +02:00
|
|
|
Add(TablePropertiesNames::kIndexKeyIsUserKey, props.index_key_is_user_key);
|
2018-08-10 01:49:45 +02:00
|
|
|
Add(TablePropertiesNames::kIndexValueIsDeltaEncoded,
|
|
|
|
props.index_value_is_delta_encoded);
|
2013-12-05 22:09:13 +01:00
|
|
|
Add(TablePropertiesNames::kNumEntries, props.num_entries);
|
2018-10-30 23:29:58 +01:00
|
|
|
Add(TablePropertiesNames::kDeletedKeys, props.num_deletions);
|
|
|
|
Add(TablePropertiesNames::kMergeOperands, props.num_merge_operands);
|
2018-06-27 05:18:43 +02:00
|
|
|
Add(TablePropertiesNames::kNumRangeDeletions, props.num_range_deletions);
|
2013-12-05 22:09:13 +01:00
|
|
|
Add(TablePropertiesNames::kNumDataBlocks, props.num_data_blocks);
|
|
|
|
Add(TablePropertiesNames::kFilterSize, props.filter_size);
|
2013-12-20 18:35:24 +01:00
|
|
|
Add(TablePropertiesNames::kFormatVersion, props.format_version);
|
|
|
|
Add(TablePropertiesNames::kFixedKeyLen, props.fixed_key_len);
|
2016-04-07 08:10:32 +02:00
|
|
|
Add(TablePropertiesNames::kColumnFamilyId, props.column_family_id);
|
2017-06-28 02:02:20 +02:00
|
|
|
Add(TablePropertiesNames::kCreationTime, props.creation_time);
|
2017-10-24 00:22:05 +02:00
|
|
|
Add(TablePropertiesNames::kOldestKeyTime, props.oldest_key_time);
|
Periodic Compactions (#5166)
Summary:
Introducing Periodic Compactions.
This feature allows all the files in a CF to be periodically compacted. It could help in catching any corruptions that could creep into the DB proactively as every file is constantly getting re-compacted. And also, of course, it helps to cleanup data older than certain threshold.
- Introduced a new option `periodic_compaction_time` to control how long a file can live without being compacted in a CF.
- This works across all levels.
- The files are put in the same level after going through the compaction. (Related files in the same level are picked up as `ExpandInputstoCleanCut` is used).
- Compaction filters, if any, are invoked as usual.
- A new table property, `file_creation_time`, is introduced to implement this feature. This property is set to the time at which the SST file was created (and that time is given by the underlying Env/OS).
This feature can be enabled on its own, or in conjunction with `ttl`. It is possible to set a different time threshold for the bottom level when used in conjunction with ttl. Since `ttl` works only on 0 to last but one levels, you could set `ttl` to, say, 1 day, and `periodic_compaction_time` to, say, 7 days. Since `ttl < periodic_compaction_time` all files in last but one levels keep getting picked up based on ttl, and almost never based on periodic_compaction_time. The files in the bottom level get picked up for compaction based on `periodic_compaction_time`.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5166
Differential Revision: D14884441
Pulled By: sagar0
fbshipit-source-id: 408426cbacb409c06386a98632dcf90bfa1bda47
2019-04-11 04:24:25 +02:00
|
|
|
if (props.file_creation_time > 0) {
|
|
|
|
Add(TablePropertiesNames::kFileCreationTime, props.file_creation_time);
|
|
|
|
}
|
2020-06-17 19:55:42 +02:00
|
|
|
Add(TablePropertiesNames::kDbId, props.db_id);
|
|
|
|
Add(TablePropertiesNames::kDbSessionId, props.db_session_id);
|
2013-12-05 22:09:13 +01:00
|
|
|
|
|
|
|
if (!props.filter_policy_name.empty()) {
|
2016-04-21 19:16:28 +02:00
|
|
|
Add(TablePropertiesNames::kFilterPolicy, props.filter_policy_name);
|
|
|
|
}
|
|
|
|
if (!props.comparator_name.empty()) {
|
|
|
|
Add(TablePropertiesNames::kComparator, props.comparator_name);
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!props.merge_operator_name.empty()) {
|
|
|
|
Add(TablePropertiesNames::kMergeOperator, props.merge_operator_name);
|
|
|
|
}
|
2016-08-26 20:46:32 +02:00
|
|
|
if (!props.prefix_extractor_name.empty()) {
|
|
|
|
Add(TablePropertiesNames::kPrefixExtractorName,
|
|
|
|
props.prefix_extractor_name);
|
|
|
|
}
|
2016-04-21 19:16:28 +02:00
|
|
|
if (!props.property_collectors_names.empty()) {
|
|
|
|
Add(TablePropertiesNames::kPropertyCollectors,
|
|
|
|
props.property_collectors_names);
|
2013-12-05 22:09:13 +01:00
|
|
|
}
|
2016-04-07 08:10:32 +02:00
|
|
|
if (!props.column_family_name.empty()) {
|
|
|
|
Add(TablePropertiesNames::kColumnFamilyName, props.column_family_name);
|
|
|
|
}
|
2016-05-12 18:47:16 +02:00
|
|
|
|
|
|
|
if (!props.compression_name.empty()) {
|
|
|
|
Add(TablePropertiesNames::kCompression, props.compression_name);
|
|
|
|
}
|
2019-04-02 23:48:52 +02:00
|
|
|
if (!props.compression_options.empty()) {
|
|
|
|
Add(TablePropertiesNames::kCompressionOptions, props.compression_options);
|
|
|
|
}
|
2013-12-05 22:09:13 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
Slice PropertyBlockBuilder::Finish() {
|
|
|
|
for (const auto& prop : props_) {
|
|
|
|
properties_block_->Add(prop.first, prop.second);
|
|
|
|
}
|
|
|
|
|
|
|
|
return properties_block_->Finish();
|
|
|
|
}
|
|
|
|
|
|
|
|
void LogPropertiesCollectionError(
|
|
|
|
Logger* info_log, const std::string& method, const std::string& name) {
|
|
|
|
assert(method == "Add" || method == "Finish");
|
|
|
|
|
|
|
|
std::string msg =
|
2014-10-30 01:55:19 +01:00
|
|
|
"Encountered error when calling TablePropertiesCollector::" +
|
2013-12-05 22:09:13 +01:00
|
|
|
method + "() with collector name: " + name;
|
2017-03-16 03:22:52 +01:00
|
|
|
ROCKS_LOG_ERROR(info_log, "%s", msg.c_str());
|
2013-12-05 22:09:13 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
bool NotifyCollectTableCollectorsOnAdd(
|
A new call back to TablePropertiesCollector to allow users know the entry is add, delete or merge
Summary:
Currently users have no idea a key is add, delete or merge from TablePropertiesCollector call back. Add a new function to add it.
Also refactor the codes so that
(1) make table property collector and internal table property collector two separate data structures with the later one now exposed
(2) table builders only receive internal table properties
Test Plan: Add cases in table_properties_collector_test to cover both of old and new ways of using TablePropertiesCollector.
Reviewers: yhchiang, igor.sugak, rven, igor
Reviewed By: rven, igor
Subscribers: meyering, yoshinorim, maykov, leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D35373
2015-04-06 19:04:30 +02:00
|
|
|
const Slice& key, const Slice& value, uint64_t file_size,
|
|
|
|
const std::vector<std::unique_ptr<IntTblPropCollector>>& collectors,
|
2013-12-05 22:09:13 +01:00
|
|
|
Logger* info_log) {
|
|
|
|
bool all_succeeded = true;
|
TablePropertiesCollectorFactory
Summary:
This diff addresses task #4296714 and rethinks how users provide us with TablePropertiesCollectors as part of Options.
Here's description of task #4296714:
I'm debugging #4295529 and noticed that our count of user properties kDeletedKeys is wrong. We're sharing one single InternalKeyPropertiesCollector with all Table Builders. In LOG Files, we're outputting number of kDeletedKeys as connected with a single table, while it's actually the total count of deleted keys since creation of the DB.
For example, this table has 3155 entries and 1391828 deleted keys.
The problem with current approach that we call methods on a single TablePropertiesCollector for all the tables we create. Even worse, we could do it from multiple threads at the same time and TablePropertiesCollector has no way of knowing which table we're calling it for.
Good part: Looks like nobody inside Facebook is using Options::table_properties_collectors. This means we should be able to painfully change the API.
In this change, I introduce TablePropertiesCollectorFactory. For every table we create, we call `CreateTablePropertiesCollector`, which creates a TablePropertiesCollector for a single table. We then use it sequentially from a single thread, which means it doesn't have to be thread-safe.
Test Plan:
Added a test in table_properties_collector_test that fails on master (build two tables, assert that kDeletedKeys count is correct for the second one).
Also, all other tests
Reviewers: sdong, dhruba, haobo, kailiu
Reviewed By: kailiu
CC: leveldb
Differential Revision: https://reviews.facebook.net/D18579
2014-05-13 21:30:55 +02:00
|
|
|
for (auto& collector : collectors) {
|
A new call back to TablePropertiesCollector to allow users know the entry is add, delete or merge
Summary:
Currently users have no idea a key is add, delete or merge from TablePropertiesCollector call back. Add a new function to add it.
Also refactor the codes so that
(1) make table property collector and internal table property collector two separate data structures with the later one now exposed
(2) table builders only receive internal table properties
Test Plan: Add cases in table_properties_collector_test to cover both of old and new ways of using TablePropertiesCollector.
Reviewers: yhchiang, igor.sugak, rven, igor
Reviewed By: rven, igor
Subscribers: meyering, yoshinorim, maykov, leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D35373
2015-04-06 19:04:30 +02:00
|
|
|
Status s = collector->InternalAdd(key, value, file_size);
|
2013-12-05 22:09:13 +01:00
|
|
|
all_succeeded = all_succeeded && s.ok();
|
|
|
|
if (!s.ok()) {
|
2014-02-05 01:21:47 +01:00
|
|
|
LogPropertiesCollectionError(info_log, "Add" /* method */,
|
|
|
|
collector->Name());
|
2013-12-05 22:09:13 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
return all_succeeded;
|
|
|
|
}
|
|
|
|
|
2019-03-18 20:07:35 +01:00
|
|
|
void NotifyCollectTableCollectorsOnBlockAdd(
|
|
|
|
const std::vector<std::unique_ptr<IntTblPropCollector>>& collectors,
|
|
|
|
const uint64_t blockRawBytes, const uint64_t blockCompressedBytesFast,
|
|
|
|
const uint64_t blockCompressedBytesSlow) {
|
|
|
|
for (auto& collector : collectors) {
|
|
|
|
collector->BlockAdd(blockRawBytes, blockCompressedBytesFast,
|
|
|
|
blockCompressedBytesSlow);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2013-12-05 22:09:13 +01:00
|
|
|
bool NotifyCollectTableCollectorsOnFinish(
|
A new call back to TablePropertiesCollector to allow users know the entry is add, delete or merge
Summary:
Currently users have no idea a key is add, delete or merge from TablePropertiesCollector call back. Add a new function to add it.
Also refactor the codes so that
(1) make table property collector and internal table property collector two separate data structures with the later one now exposed
(2) table builders only receive internal table properties
Test Plan: Add cases in table_properties_collector_test to cover both of old and new ways of using TablePropertiesCollector.
Reviewers: yhchiang, igor.sugak, rven, igor
Reviewed By: rven, igor
Subscribers: meyering, yoshinorim, maykov, leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D35373
2015-04-06 19:04:30 +02:00
|
|
|
const std::vector<std::unique_ptr<IntTblPropCollector>>& collectors,
|
TablePropertiesCollectorFactory
Summary:
This diff addresses task #4296714 and rethinks how users provide us with TablePropertiesCollectors as part of Options.
Here's description of task #4296714:
I'm debugging #4295529 and noticed that our count of user properties kDeletedKeys is wrong. We're sharing one single InternalKeyPropertiesCollector with all Table Builders. In LOG Files, we're outputting number of kDeletedKeys as connected with a single table, while it's actually the total count of deleted keys since creation of the DB.
For example, this table has 3155 entries and 1391828 deleted keys.
The problem with current approach that we call methods on a single TablePropertiesCollector for all the tables we create. Even worse, we could do it from multiple threads at the same time and TablePropertiesCollector has no way of knowing which table we're calling it for.
Good part: Looks like nobody inside Facebook is using Options::table_properties_collectors. This means we should be able to painfully change the API.
In this change, I introduce TablePropertiesCollectorFactory. For every table we create, we call `CreateTablePropertiesCollector`, which creates a TablePropertiesCollector for a single table. We then use it sequentially from a single thread, which means it doesn't have to be thread-safe.
Test Plan:
Added a test in table_properties_collector_test that fails on master (build two tables, assert that kDeletedKeys count is correct for the second one).
Also, all other tests
Reviewers: sdong, dhruba, haobo, kailiu
Reviewed By: kailiu
CC: leveldb
Differential Revision: https://reviews.facebook.net/D18579
2014-05-13 21:30:55 +02:00
|
|
|
Logger* info_log, PropertyBlockBuilder* builder) {
|
2013-12-05 22:09:13 +01:00
|
|
|
bool all_succeeded = true;
|
TablePropertiesCollectorFactory
Summary:
This diff addresses task #4296714 and rethinks how users provide us with TablePropertiesCollectors as part of Options.
Here's description of task #4296714:
I'm debugging #4295529 and noticed that our count of user properties kDeletedKeys is wrong. We're sharing one single InternalKeyPropertiesCollector with all Table Builders. In LOG Files, we're outputting number of kDeletedKeys as connected with a single table, while it's actually the total count of deleted keys since creation of the DB.
For example, this table has 3155 entries and 1391828 deleted keys.
The problem with current approach that we call methods on a single TablePropertiesCollector for all the tables we create. Even worse, we could do it from multiple threads at the same time and TablePropertiesCollector has no way of knowing which table we're calling it for.
Good part: Looks like nobody inside Facebook is using Options::table_properties_collectors. This means we should be able to painfully change the API.
In this change, I introduce TablePropertiesCollectorFactory. For every table we create, we call `CreateTablePropertiesCollector`, which creates a TablePropertiesCollector for a single table. We then use it sequentially from a single thread, which means it doesn't have to be thread-safe.
Test Plan:
Added a test in table_properties_collector_test that fails on master (build two tables, assert that kDeletedKeys count is correct for the second one).
Also, all other tests
Reviewers: sdong, dhruba, haobo, kailiu
Reviewed By: kailiu
CC: leveldb
Differential Revision: https://reviews.facebook.net/D18579
2014-05-13 21:30:55 +02:00
|
|
|
for (auto& collector : collectors) {
|
2013-12-05 22:09:13 +01:00
|
|
|
UserCollectedProperties user_collected_properties;
|
|
|
|
Status s = collector->Finish(&user_collected_properties);
|
|
|
|
|
|
|
|
all_succeeded = all_succeeded && s.ok();
|
|
|
|
if (!s.ok()) {
|
2014-02-05 01:21:47 +01:00
|
|
|
LogPropertiesCollectionError(info_log, "Finish" /* method */,
|
|
|
|
collector->Name());
|
2013-12-05 22:09:13 +01:00
|
|
|
} else {
|
|
|
|
builder->Add(user_collected_properties);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return all_succeeded;
|
|
|
|
}
|
|
|
|
|
2020-06-29 23:51:57 +02:00
|
|
|
Status ReadProperties(const ReadOptions& read_options,
|
|
|
|
const Slice& handle_value, RandomAccessFileReader* file,
|
2017-08-11 20:59:13 +02:00
|
|
|
FilePrefetchBuffer* prefetch_buffer, const Footer& footer,
|
|
|
|
const ImmutableCFOptions& ioptions,
|
2019-02-11 20:37:07 +01:00
|
|
|
TableProperties** table_properties, bool verify_checksum,
|
|
|
|
BlockHandle* ret_block_handle,
|
|
|
|
CacheAllocationPtr* verification_buf,
|
2018-11-29 02:58:08 +01:00
|
|
|
bool /*compression_type_missing*/,
|
|
|
|
MemoryAllocator* memory_allocator) {
|
2013-12-06 01:51:26 +01:00
|
|
|
assert(table_properties);
|
|
|
|
|
|
|
|
Slice v = handle_value;
|
|
|
|
BlockHandle handle;
|
|
|
|
if (!handle.DecodeFrom(&v).ok()) {
|
|
|
|
return Status::InvalidArgument("Failed to decode properties block handle");
|
|
|
|
}
|
|
|
|
|
|
|
|
BlockContents block_contents;
|
2014-08-16 00:05:09 +02:00
|
|
|
Status s;
|
2017-12-12 00:16:37 +01:00
|
|
|
PersistentCacheOptions cache_options;
|
2020-06-29 23:51:57 +02:00
|
|
|
ReadOptions ro = read_options;
|
|
|
|
ro.verify_checksums = verify_checksum;
|
|
|
|
|
|
|
|
BlockFetcher block_fetcher(file, prefetch_buffer, footer, ro, handle,
|
|
|
|
&block_contents, ioptions, false /* decompress */,
|
|
|
|
false /*maybe_compressed*/, BlockType::kProperties,
|
|
|
|
UncompressionDict::GetEmptyDict(), cache_options,
|
|
|
|
memory_allocator);
|
2017-12-12 00:16:37 +01:00
|
|
|
s = block_fetcher.ReadBlockContents();
|
2018-11-14 02:00:49 +01:00
|
|
|
// property block is never compressed. Need to add uncompress logic if we are
|
|
|
|
// to compress it..
|
2013-12-06 01:51:26 +01:00
|
|
|
|
|
|
|
if (!s.ok()) {
|
|
|
|
return s;
|
|
|
|
}
|
|
|
|
|
2020-02-26 00:29:17 +01:00
|
|
|
Block properties_block(std::move(block_contents));
|
2018-07-16 18:58:58 +02:00
|
|
|
DataBlockIter iter;
|
Add an option to put first key of each sst block in the index (#5289)
Summary:
The first key is used to defer reading the data block until this file gets to the top of merging iterator's heap. For short range scans, most files never make it to the top of the heap, so this change can reduce read amplification by a lot sometimes.
Consider the following workload. There are a few data streams (we'll be calling them "logs"), each stream consisting of a sequence of blobs (we'll be calling them "records"). Each record is identified by log ID and a sequence number within the log. RocksDB key is concatenation of log ID and sequence number (big endian). Reads are mostly relatively short range scans, each within a single log. Writes are mostly sequential for each log, but writes to different logs are randomly interleaved. Compactions are disabled; instead, when we accumulate a few tens of sst files, we create a new column family and start writing to it.
So, a typical sst file consists of a few ranges of blocks, each range corresponding to one log ID (we use FlushBlockPolicy to cut blocks at log boundaries). A typical read would go like this. First, iterator Seek() reads one block from each sst file. Then a series of Next()s move through one sst file (since writes to each log are mostly sequential) until the subiterator reaches the end of this log in this sst file; then Next() switches to the next sst file and reads sequentially from that, and so on. Often a range scan will only return records from a small number of blocks in small number of sst files; in this case, the cost of initial Seek() reading one block from each file may be bigger than the cost of reading the actually useful blocks.
Neither iterate_upper_bound nor bloom filters can prevent reading one block from each file in Seek(). But this PR can: if the index contains first key from each block, we don't have to read the block until this block actually makes it to the top of merging iterator's heap, so for short range scans we won't read any blocks from most of the sst files.
This PR does the deferred block loading inside value() call. This is not ideal: there's no good way to report an IO error from inside value(). As discussed with siying offline, it would probably be better to change InternalIterator's interface to explicitly fetch deferred value and get status. I'll do it in a separate PR.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5289
Differential Revision: D15256423
Pulled By: al13n321
fbshipit-source-id: 750e4c39ce88e8d41662f701cf6275d9388ba46a
2019-06-25 05:50:35 +02:00
|
|
|
properties_block.NewDataIterator(BytewiseComparator(), BytewiseComparator(),
|
2020-02-26 00:29:17 +01:00
|
|
|
kDisableGlobalSequenceNumber, &iter);
|
2013-12-06 01:51:26 +01:00
|
|
|
|
2014-02-08 04:26:49 +01:00
|
|
|
auto new_table_properties = new TableProperties();
|
2013-12-06 01:51:26 +01:00
|
|
|
// All pre-defined properties of type uint64_t
|
|
|
|
std::unordered_map<std::string, uint64_t*> predefined_uint64_properties = {
|
2014-02-08 04:26:49 +01:00
|
|
|
{TablePropertiesNames::kDataSize, &new_table_properties->data_size},
|
|
|
|
{TablePropertiesNames::kIndexSize, &new_table_properties->index_size},
|
2017-06-13 19:59:22 +02:00
|
|
|
{TablePropertiesNames::kIndexPartitions,
|
|
|
|
&new_table_properties->index_partitions},
|
|
|
|
{TablePropertiesNames::kTopLevelIndexSize,
|
|
|
|
&new_table_properties->top_level_index_size},
|
2018-05-26 03:41:31 +02:00
|
|
|
{TablePropertiesNames::kIndexKeyIsUserKey,
|
|
|
|
&new_table_properties->index_key_is_user_key},
|
2018-08-10 01:49:45 +02:00
|
|
|
{TablePropertiesNames::kIndexValueIsDeltaEncoded,
|
|
|
|
&new_table_properties->index_value_is_delta_encoded},
|
2014-02-08 04:26:49 +01:00
|
|
|
{TablePropertiesNames::kFilterSize, &new_table_properties->filter_size},
|
|
|
|
{TablePropertiesNames::kRawKeySize, &new_table_properties->raw_key_size},
|
|
|
|
{TablePropertiesNames::kRawValueSize,
|
|
|
|
&new_table_properties->raw_value_size},
|
2014-02-05 01:21:47 +01:00
|
|
|
{TablePropertiesNames::kNumDataBlocks,
|
2014-02-08 04:26:49 +01:00
|
|
|
&new_table_properties->num_data_blocks},
|
|
|
|
{TablePropertiesNames::kNumEntries, &new_table_properties->num_entries},
|
2018-10-30 23:29:58 +01:00
|
|
|
{TablePropertiesNames::kDeletedKeys,
|
|
|
|
&new_table_properties->num_deletions},
|
|
|
|
{TablePropertiesNames::kMergeOperands,
|
|
|
|
&new_table_properties->num_merge_operands},
|
2018-06-27 05:18:43 +02:00
|
|
|
{TablePropertiesNames::kNumRangeDeletions,
|
|
|
|
&new_table_properties->num_range_deletions},
|
2014-02-08 04:26:49 +01:00
|
|
|
{TablePropertiesNames::kFormatVersion,
|
|
|
|
&new_table_properties->format_version},
|
|
|
|
{TablePropertiesNames::kFixedKeyLen,
|
2016-04-07 08:10:32 +02:00
|
|
|
&new_table_properties->fixed_key_len},
|
|
|
|
{TablePropertiesNames::kColumnFamilyId,
|
|
|
|
&new_table_properties->column_family_id},
|
2017-06-28 02:02:20 +02:00
|
|
|
{TablePropertiesNames::kCreationTime,
|
|
|
|
&new_table_properties->creation_time},
|
2017-10-24 00:22:05 +02:00
|
|
|
{TablePropertiesNames::kOldestKeyTime,
|
|
|
|
&new_table_properties->oldest_key_time},
|
Periodic Compactions (#5166)
Summary:
Introducing Periodic Compactions.
This feature allows all the files in a CF to be periodically compacted. It could help in catching any corruptions that could creep into the DB proactively as every file is constantly getting re-compacted. And also, of course, it helps to cleanup data older than certain threshold.
- Introduced a new option `periodic_compaction_time` to control how long a file can live without being compacted in a CF.
- This works across all levels.
- The files are put in the same level after going through the compaction. (Related files in the same level are picked up as `ExpandInputstoCleanCut` is used).
- Compaction filters, if any, are invoked as usual.
- A new table property, `file_creation_time`, is introduced to implement this feature. This property is set to the time at which the SST file was created (and that time is given by the underlying Env/OS).
This feature can be enabled on its own, or in conjunction with `ttl`. It is possible to set a different time threshold for the bottom level when used in conjunction with ttl. Since `ttl` works only on 0 to last but one levels, you could set `ttl` to, say, 1 day, and `periodic_compaction_time` to, say, 7 days. Since `ttl < periodic_compaction_time` all files in last but one levels keep getting picked up based on ttl, and almost never based on periodic_compaction_time. The files in the bottom level get picked up for compaction based on `periodic_compaction_time`.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5166
Differential Revision: D14884441
Pulled By: sagar0
fbshipit-source-id: 408426cbacb409c06386a98632dcf90bfa1bda47
2019-04-11 04:24:25 +02:00
|
|
|
{TablePropertiesNames::kFileCreationTime,
|
|
|
|
&new_table_properties->file_creation_time},
|
2016-04-07 08:10:32 +02:00
|
|
|
};
|
2013-12-06 01:51:26 +01:00
|
|
|
|
|
|
|
std::string last_key;
|
2019-02-11 20:37:07 +01:00
|
|
|
for (iter.SeekToFirstOrReport(); iter.Valid(); iter.NextOrReport()) {
|
2016-10-19 01:59:37 +02:00
|
|
|
s = iter.status();
|
2013-12-06 01:51:26 +01:00
|
|
|
if (!s.ok()) {
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
2016-10-19 01:59:37 +02:00
|
|
|
auto key = iter.key().ToString();
|
2019-02-11 20:37:07 +01:00
|
|
|
// properties block should be strictly sorted with no duplicate key.
|
|
|
|
if (!last_key.empty() &&
|
|
|
|
BytewiseComparator()->Compare(key, last_key) <= 0) {
|
|
|
|
s = Status::Corruption("properties unsorted");
|
|
|
|
break;
|
|
|
|
}
|
2013-12-06 01:51:26 +01:00
|
|
|
last_key = key;
|
|
|
|
|
2016-10-19 01:59:37 +02:00
|
|
|
auto raw_val = iter.value();
|
2013-12-06 01:51:26 +01:00
|
|
|
auto pos = predefined_uint64_properties.find(key);
|
|
|
|
|
2016-10-19 01:59:37 +02:00
|
|
|
new_table_properties->properties_offsets.insert(
|
|
|
|
{key, handle.offset() + iter.ValueOffset()});
|
|
|
|
|
2013-12-06 01:51:26 +01:00
|
|
|
if (pos != predefined_uint64_properties.end()) {
|
2018-10-30 23:29:58 +01:00
|
|
|
if (key == TablePropertiesNames::kDeletedKeys ||
|
|
|
|
key == TablePropertiesNames::kMergeOperands) {
|
|
|
|
// Insert in user-collected properties for API backwards compatibility
|
|
|
|
new_table_properties->user_collected_properties.insert(
|
|
|
|
{key, raw_val.ToString()});
|
|
|
|
}
|
2013-12-06 01:51:26 +01:00
|
|
|
// handle predefined rocksdb properties
|
|
|
|
uint64_t val;
|
|
|
|
if (!GetVarint64(&raw_val, &val)) {
|
|
|
|
// skip malformed value
|
|
|
|
auto error_msg =
|
2014-10-30 01:55:19 +01:00
|
|
|
"Detect malformed value in properties meta-block:"
|
2013-12-06 01:51:26 +01:00
|
|
|
"\tkey: " + key + "\tval: " + raw_val.ToString();
|
2017-03-16 03:22:52 +01:00
|
|
|
ROCKS_LOG_ERROR(ioptions.info_log, "%s", error_msg.c_str());
|
2013-12-06 01:51:26 +01:00
|
|
|
continue;
|
|
|
|
}
|
|
|
|
*(pos->second) = val;
|
2020-06-17 19:55:42 +02:00
|
|
|
} else if (key == TablePropertiesNames::kDbId) {
|
|
|
|
new_table_properties->db_id = raw_val.ToString();
|
|
|
|
} else if (key == TablePropertiesNames::kDbSessionId) {
|
|
|
|
new_table_properties->db_session_id = raw_val.ToString();
|
2013-12-06 01:51:26 +01:00
|
|
|
} else if (key == TablePropertiesNames::kFilterPolicy) {
|
2014-02-08 04:26:49 +01:00
|
|
|
new_table_properties->filter_policy_name = raw_val.ToString();
|
2016-04-07 08:10:32 +02:00
|
|
|
} else if (key == TablePropertiesNames::kColumnFamilyName) {
|
|
|
|
new_table_properties->column_family_name = raw_val.ToString();
|
2016-04-21 19:16:28 +02:00
|
|
|
} else if (key == TablePropertiesNames::kComparator) {
|
|
|
|
new_table_properties->comparator_name = raw_val.ToString();
|
|
|
|
} else if (key == TablePropertiesNames::kMergeOperator) {
|
|
|
|
new_table_properties->merge_operator_name = raw_val.ToString();
|
2016-08-26 20:46:32 +02:00
|
|
|
} else if (key == TablePropertiesNames::kPrefixExtractorName) {
|
|
|
|
new_table_properties->prefix_extractor_name = raw_val.ToString();
|
2016-04-21 19:16:28 +02:00
|
|
|
} else if (key == TablePropertiesNames::kPropertyCollectors) {
|
|
|
|
new_table_properties->property_collectors_names = raw_val.ToString();
|
2016-05-12 18:47:16 +02:00
|
|
|
} else if (key == TablePropertiesNames::kCompression) {
|
|
|
|
new_table_properties->compression_name = raw_val.ToString();
|
2019-04-02 23:48:52 +02:00
|
|
|
} else if (key == TablePropertiesNames::kCompressionOptions) {
|
|
|
|
new_table_properties->compression_options = raw_val.ToString();
|
2013-12-06 01:51:26 +01:00
|
|
|
} else {
|
|
|
|
// handle user-collected properties
|
2014-02-08 04:26:49 +01:00
|
|
|
new_table_properties->user_collected_properties.insert(
|
2014-02-05 01:21:47 +01:00
|
|
|
{key, raw_val.ToString()});
|
2013-12-06 01:51:26 +01:00
|
|
|
}
|
|
|
|
}
|
2014-02-08 04:26:49 +01:00
|
|
|
if (s.ok()) {
|
|
|
|
*table_properties = new_table_properties;
|
2019-02-11 20:37:07 +01:00
|
|
|
if (ret_block_handle != nullptr) {
|
|
|
|
*ret_block_handle = handle;
|
|
|
|
}
|
|
|
|
if (verification_buf != nullptr) {
|
2019-04-23 00:59:16 +02:00
|
|
|
size_t len = static_cast<size_t>(handle.size() + kBlockTrailerSize);
|
2020-02-20 21:07:53 +01:00
|
|
|
*verification_buf =
|
|
|
|
ROCKSDB_NAMESPACE::AllocateBlock(len, memory_allocator);
|
2019-02-11 20:37:07 +01:00
|
|
|
if (verification_buf->get() != nullptr) {
|
|
|
|
memcpy(verification_buf->get(), block_contents.data.data(), len);
|
|
|
|
}
|
|
|
|
}
|
2014-02-08 04:26:49 +01:00
|
|
|
} else {
|
|
|
|
delete new_table_properties;
|
|
|
|
}
|
2013-12-06 01:51:26 +01:00
|
|
|
|
|
|
|
return s;
|
|
|
|
}
|
|
|
|
|
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
2015-07-18 01:16:11 +02:00
|
|
|
Status ReadTableProperties(RandomAccessFileReader* file, uint64_t file_size,
|
2016-07-19 18:44:03 +02:00
|
|
|
uint64_t table_magic_number,
|
2018-11-29 02:58:08 +01:00
|
|
|
const ImmutableCFOptions& ioptions,
|
2018-06-16 04:24:21 +02:00
|
|
|
TableProperties** properties,
|
2018-11-29 02:58:08 +01:00
|
|
|
bool compression_type_missing,
|
2020-05-13 03:21:32 +02:00
|
|
|
MemoryAllocator* memory_allocator,
|
|
|
|
FilePrefetchBuffer* prefetch_buffer) {
|
2013-12-06 01:51:26 +01:00
|
|
|
// -- Read metaindex block
|
2015-01-13 23:33:04 +01:00
|
|
|
Footer footer;
|
2020-06-29 23:51:57 +02:00
|
|
|
IOOptions opts;
|
|
|
|
auto s = ReadFooterFromFile(opts, file, prefetch_buffer, file_size, &footer,
|
2020-05-13 03:21:32 +02:00
|
|
|
table_magic_number);
|
2013-12-06 01:51:26 +01:00
|
|
|
if (!s.ok()) {
|
|
|
|
return s;
|
|
|
|
}
|
|
|
|
|
|
|
|
auto metaindex_handle = footer.metaindex_handle();
|
|
|
|
BlockContents metaindex_contents;
|
2014-02-06 05:56:14 +01:00
|
|
|
ReadOptions read_options;
|
|
|
|
read_options.verify_checksums = false;
|
2017-12-12 00:16:37 +01:00
|
|
|
PersistentCacheOptions cache_options;
|
|
|
|
|
2019-01-24 03:11:08 +01:00
|
|
|
BlockFetcher block_fetcher(
|
2020-05-13 03:21:32 +02:00
|
|
|
file, prefetch_buffer, footer, read_options, metaindex_handle,
|
|
|
|
&metaindex_contents, ioptions, false /* decompress */,
|
2019-06-19 04:00:03 +02:00
|
|
|
false /*maybe_compressed*/, BlockType::kMetaIndex,
|
|
|
|
UncompressionDict::GetEmptyDict(), cache_options, memory_allocator);
|
2017-12-12 00:16:37 +01:00
|
|
|
s = block_fetcher.ReadBlockContents();
|
2013-12-06 01:51:26 +01:00
|
|
|
if (!s.ok()) {
|
|
|
|
return s;
|
|
|
|
}
|
2018-11-14 02:00:49 +01:00
|
|
|
// property blocks are never compressed. Need to add uncompress logic if we
|
|
|
|
// are to compress it.
|
2020-02-26 00:29:17 +01:00
|
|
|
Block metaindex_block(std::move(metaindex_contents));
|
Add an option to put first key of each sst block in the index (#5289)
Summary:
The first key is used to defer reading the data block until this file gets to the top of merging iterator's heap. For short range scans, most files never make it to the top of the heap, so this change can reduce read amplification by a lot sometimes.
Consider the following workload. There are a few data streams (we'll be calling them "logs"), each stream consisting of a sequence of blobs (we'll be calling them "records"). Each record is identified by log ID and a sequence number within the log. RocksDB key is concatenation of log ID and sequence number (big endian). Reads are mostly relatively short range scans, each within a single log. Writes are mostly sequential for each log, but writes to different logs are randomly interleaved. Compactions are disabled; instead, when we accumulate a few tens of sst files, we create a new column family and start writing to it.
So, a typical sst file consists of a few ranges of blocks, each range corresponding to one log ID (we use FlushBlockPolicy to cut blocks at log boundaries). A typical read would go like this. First, iterator Seek() reads one block from each sst file. Then a series of Next()s move through one sst file (since writes to each log are mostly sequential) until the subiterator reaches the end of this log in this sst file; then Next() switches to the next sst file and reads sequentially from that, and so on. Often a range scan will only return records from a small number of blocks in small number of sst files; in this case, the cost of initial Seek() reading one block from each file may be bigger than the cost of reading the actually useful blocks.
Neither iterate_upper_bound nor bloom filters can prevent reading one block from each file in Seek(). But this PR can: if the index contains first key from each block, we don't have to read the block until this block actually makes it to the top of merging iterator's heap, so for short range scans we won't read any blocks from most of the sst files.
This PR does the deferred block loading inside value() call. This is not ideal: there's no good way to report an IO error from inside value(). As discussed with siying offline, it would probably be better to change InternalIterator's interface to explicitly fetch deferred value and get status. I'll do it in a separate PR.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5289
Differential Revision: D15256423
Pulled By: al13n321
fbshipit-source-id: 750e4c39ce88e8d41662f701cf6275d9388ba46a
2019-06-25 05:50:35 +02:00
|
|
|
std::unique_ptr<InternalIterator> meta_iter(metaindex_block.NewDataIterator(
|
2020-02-26 00:29:17 +01:00
|
|
|
BytewiseComparator(), BytewiseComparator(),
|
|
|
|
kDisableGlobalSequenceNumber));
|
2013-12-06 01:51:26 +01:00
|
|
|
|
|
|
|
// -- Read property block
|
2014-04-22 02:49:47 +02:00
|
|
|
bool found_properties_block = true;
|
|
|
|
s = SeekToPropertiesBlock(meta_iter.get(), &found_properties_block);
|
|
|
|
if (!s.ok()) {
|
|
|
|
return s;
|
|
|
|
}
|
|
|
|
|
2013-12-06 01:51:26 +01:00
|
|
|
TableProperties table_properties;
|
2014-04-22 02:49:47 +02:00
|
|
|
if (found_properties_block == true) {
|
2020-06-29 23:51:57 +02:00
|
|
|
s = ReadProperties(
|
|
|
|
read_options, meta_iter->value(), file, prefetch_buffer, footer,
|
|
|
|
ioptions, properties, false /* verify_checksum */,
|
|
|
|
nullptr /* ret_block_hanel */, nullptr /* ret_block_contents */,
|
|
|
|
compression_type_missing, memory_allocator);
|
2013-12-06 01:51:26 +01:00
|
|
|
} else {
|
2014-08-15 21:17:44 +02:00
|
|
|
s = Status::NotFound();
|
2013-12-06 01:51:26 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
return s;
|
|
|
|
}
|
|
|
|
|
2015-10-13 00:06:38 +02:00
|
|
|
Status FindMetaBlock(InternalIterator* meta_index_iter,
|
2014-05-15 23:09:03 +02:00
|
|
|
const std::string& meta_block_name,
|
|
|
|
BlockHandle* block_handle) {
|
|
|
|
meta_index_iter->Seek(meta_block_name);
|
|
|
|
if (meta_index_iter->status().ok() && meta_index_iter->Valid() &&
|
|
|
|
meta_index_iter->key() == meta_block_name) {
|
|
|
|
Slice v = meta_index_iter->value();
|
|
|
|
return block_handle->DecodeFrom(&v);
|
|
|
|
} else {
|
|
|
|
return Status::Corruption("Cannot find the meta block", meta_block_name);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
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
2015-07-18 01:16:11 +02:00
|
|
|
Status FindMetaBlock(RandomAccessFileReader* file, uint64_t file_size,
|
2016-07-19 18:44:03 +02:00
|
|
|
uint64_t table_magic_number,
|
2018-11-14 02:00:49 +01:00
|
|
|
const ImmutableCFOptions& ioptions,
|
2014-07-19 01:58:13 +02:00
|
|
|
const std::string& meta_block_name,
|
2018-06-16 04:24:21 +02:00
|
|
|
BlockHandle* block_handle,
|
2018-11-29 02:58:08 +01:00
|
|
|
bool /*compression_type_missing*/,
|
|
|
|
MemoryAllocator* memory_allocator) {
|
2015-01-13 23:33:04 +01:00
|
|
|
Footer footer;
|
2020-06-29 23:51:57 +02:00
|
|
|
IOOptions opts;
|
|
|
|
auto s = ReadFooterFromFile(opts, file, nullptr /* prefetch_buffer */,
|
|
|
|
file_size, &footer, table_magic_number);
|
2014-07-19 01:58:13 +02:00
|
|
|
if (!s.ok()) {
|
|
|
|
return s;
|
|
|
|
}
|
|
|
|
|
|
|
|
auto metaindex_handle = footer.metaindex_handle();
|
|
|
|
BlockContents metaindex_contents;
|
|
|
|
ReadOptions read_options;
|
|
|
|
read_options.verify_checksums = false;
|
2017-12-12 00:16:37 +01:00
|
|
|
PersistentCacheOptions cache_options;
|
|
|
|
BlockFetcher block_fetcher(
|
|
|
|
file, nullptr /* prefetch_buffer */, footer, read_options,
|
|
|
|
metaindex_handle, &metaindex_contents, ioptions,
|
2018-11-29 02:58:08 +01:00
|
|
|
false /* do decompression */, false /*maybe_compressed*/,
|
2019-06-19 04:00:03 +02:00
|
|
|
BlockType::kMetaIndex, UncompressionDict::GetEmptyDict(), cache_options,
|
|
|
|
memory_allocator);
|
2017-12-12 00:16:37 +01:00
|
|
|
s = block_fetcher.ReadBlockContents();
|
2014-07-19 01:58:13 +02:00
|
|
|
if (!s.ok()) {
|
|
|
|
return s;
|
|
|
|
}
|
2018-11-14 02:00:49 +01:00
|
|
|
// meta blocks are never compressed. Need to add uncompress logic if we are to
|
|
|
|
// compress it.
|
2020-02-26 00:29:17 +01:00
|
|
|
Block metaindex_block(std::move(metaindex_contents));
|
2014-07-19 01:58:13 +02:00
|
|
|
|
2015-10-13 00:06:38 +02:00
|
|
|
std::unique_ptr<InternalIterator> meta_iter;
|
2020-02-26 00:29:17 +01:00
|
|
|
meta_iter.reset(metaindex_block.NewDataIterator(
|
|
|
|
BytewiseComparator(), BytewiseComparator(),
|
|
|
|
kDisableGlobalSequenceNumber));
|
2014-07-19 01:58:13 +02:00
|
|
|
|
|
|
|
return FindMetaBlock(meta_iter.get(), meta_block_name, block_handle);
|
|
|
|
}
|
|
|
|
|
2017-08-11 20:59:13 +02:00
|
|
|
Status ReadMetaBlock(RandomAccessFileReader* file,
|
|
|
|
FilePrefetchBuffer* prefetch_buffer, uint64_t file_size,
|
2016-07-19 18:44:03 +02:00
|
|
|
uint64_t table_magic_number,
|
2017-08-11 20:59:13 +02:00
|
|
|
const ImmutableCFOptions& ioptions,
|
2019-06-19 04:00:03 +02:00
|
|
|
const std::string& meta_block_name, BlockType block_type,
|
2018-11-29 02:58:08 +01:00
|
|
|
BlockContents* contents, bool /*compression_type_missing*/,
|
|
|
|
MemoryAllocator* memory_allocator) {
|
2014-08-16 00:05:09 +02:00
|
|
|
Status status;
|
2015-01-13 23:33:04 +01:00
|
|
|
Footer footer;
|
2020-06-29 23:51:57 +02:00
|
|
|
IOOptions opts;
|
|
|
|
status = ReadFooterFromFile(opts, file, prefetch_buffer, file_size, &footer,
|
2017-08-11 20:59:13 +02:00
|
|
|
table_magic_number);
|
2014-09-18 00:08:19 +02:00
|
|
|
if (!status.ok()) {
|
|
|
|
return status;
|
|
|
|
}
|
2014-07-19 01:58:13 +02:00
|
|
|
|
|
|
|
// Reading metaindex block
|
|
|
|
auto metaindex_handle = footer.metaindex_handle();
|
|
|
|
BlockContents metaindex_contents;
|
|
|
|
ReadOptions read_options;
|
|
|
|
read_options.verify_checksums = false;
|
2017-12-12 00:16:37 +01:00
|
|
|
PersistentCacheOptions cache_options;
|
|
|
|
|
2019-06-19 23:07:36 +02:00
|
|
|
BlockFetcher block_fetcher(
|
|
|
|
file, prefetch_buffer, footer, read_options, metaindex_handle,
|
|
|
|
&metaindex_contents, ioptions, false /* decompress */,
|
|
|
|
false /*maybe_compressed*/, BlockType::kMetaIndex,
|
|
|
|
UncompressionDict::GetEmptyDict(), cache_options, memory_allocator);
|
2017-12-12 00:16:37 +01:00
|
|
|
status = block_fetcher.ReadBlockContents();
|
2014-09-18 00:08:19 +02:00
|
|
|
if (!status.ok()) {
|
|
|
|
return status;
|
|
|
|
}
|
2018-11-14 02:00:49 +01:00
|
|
|
// meta block is never compressed. Need to add uncompress logic if we are to
|
|
|
|
// compress it.
|
2014-07-19 01:58:13 +02:00
|
|
|
|
|
|
|
// Finding metablock
|
2020-02-26 00:29:17 +01:00
|
|
|
Block metaindex_block(std::move(metaindex_contents));
|
2014-07-19 01:58:13 +02:00
|
|
|
|
2015-10-13 00:06:38 +02:00
|
|
|
std::unique_ptr<InternalIterator> meta_iter;
|
2020-02-26 00:29:17 +01:00
|
|
|
meta_iter.reset(metaindex_block.NewDataIterator(
|
|
|
|
BytewiseComparator(), BytewiseComparator(),
|
|
|
|
kDisableGlobalSequenceNumber));
|
2014-07-19 01:58:13 +02:00
|
|
|
|
|
|
|
BlockHandle block_handle;
|
2014-08-16 00:05:09 +02:00
|
|
|
status = FindMetaBlock(meta_iter.get(), meta_block_name, &block_handle);
|
2014-07-19 01:58:13 +02:00
|
|
|
|
2014-08-16 00:05:09 +02:00
|
|
|
if (!status.ok()) {
|
|
|
|
return status;
|
2014-07-19 01:58:13 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
// Reading metablock
|
2017-12-12 00:16:37 +01:00
|
|
|
BlockFetcher block_fetcher2(
|
|
|
|
file, prefetch_buffer, footer, read_options, block_handle, contents,
|
2019-06-19 04:00:03 +02:00
|
|
|
ioptions, false /* decompress */, false /*maybe_compressed*/, block_type,
|
2019-01-24 03:11:08 +01:00
|
|
|
UncompressionDict::GetEmptyDict(), cache_options, memory_allocator);
|
2017-12-12 00:16:37 +01:00
|
|
|
return block_fetcher2.ReadBlockContents();
|
2014-07-19 01:58:13 +02:00
|
|
|
}
|
|
|
|
|
2020-02-20 21:07:53 +01:00
|
|
|
} // namespace ROCKSDB_NAMESPACE
|