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/db/db_table_properties_test.cc

625 lines
21 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).
//
// Copyright (c) 2011 The LevelDB Authors. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file. See the AUTHORS file for names of contributors.
Account memory of big memory users in BlockBasedTable in global memory limit (#9748) Summary: **Context:** Through heap profiling, we discovered that `BlockBasedTableReader` objects can accumulate and lead to high memory usage (e.g, `max_open_file = -1`). These memories are currently not saved, not tracked, not constrained and not cache evict-able. As a first step to improve this, similar to https://github.com/facebook/rocksdb/pull/8428, this PR is to track an estimate of `BlockBasedTableReader` object's memory in block cache and fail future creation if the memory usage exceeds the available space of cache at the time of creation. **Summary:** - Approximate big memory users (`BlockBasedTable::Rep` and `TableProperties` )' memory usage in addition to the existing estimated ones (filter block/index block/un-compression dictionary) - Charge all of these memory usages to block cache on `BlockBasedTable::Open()` and release them on `~BlockBasedTable()` as there is no memory usage fluctuation of concern in between - Refactor on CacheReservationManager (and its call-sites) to add concurrent support for BlockBasedTable used in this PR. Pull Request resolved: https://github.com/facebook/rocksdb/pull/9748 Test Plan: - New unit tests - db bench: `OpenDb` : **-0.52% in ms** - Setup `./db_bench -benchmarks=fillseq -db=/dev/shm/testdb -disable_auto_compactions=1 -write_buffer_size=1048576` - Repeated run with pre-change w/o feature and post-change with feature, benchmark `OpenDb`: `./db_bench -benchmarks=readrandom -use_existing_db=1 -db=/dev/shm/testdb -reserve_table_reader_memory=true (remove this when running w/o feature) -file_opening_threads=3 -open_files=-1 -report_open_timing=true| egrep 'OpenDb:'` #-run | (feature-off) avg milliseconds | std milliseconds | (feature-on) avg milliseconds | std milliseconds | change (%) -- | -- | -- | -- | -- | -- 10 | 11.4018 | 5.95173 | 9.47788 | 1.57538 | -16.87382694 20 | 9.23746 | 0.841053 | 9.32377 | 1.14074 | 0.9343477536 40 | 9.0876 | 0.671129 | 9.35053 | 1.11713 | 2.893283155 80 | 9.72514 | 2.28459 | 9.52013 | 1.0894 | -2.108041632 160 | 9.74677 | 0.991234 | 9.84743 | 1.73396 | 1.032752389 320 | 10.7297 | 5.11555 | 10.547 | 1.97692 | **-1.70275031** 640 | 11.7092 | 2.36565 | 11.7869 | 2.69377 | **0.6635807741** - db bench on write with cost to cache in WriteBufferManager (just in case this PR's CRM refactoring accidentally slows down anything in WBM) : `fillseq` : **+0.54% in micros/op** `./db_bench -benchmarks=fillseq -db=/dev/shm/testdb -disable_auto_compactions=1 -cost_write_buffer_to_cache=true -write_buffer_size=10000000000 | egrep 'fillseq'` #-run | (pre-PR) avg micros/op | std micros/op | (post-PR) avg micros/op | std micros/op | change (%) -- | -- | -- | -- | -- | -- 10 | 6.15 | 0.260187 | 6.289 | 0.371192 | 2.260162602 20 | 7.28025 | 0.465402 | 7.37255 | 0.451256 | 1.267813605 40 | 7.06312 | 0.490654 | 7.13803 | 0.478676 | **1.060579461** 80 | 7.14035 | 0.972831 | 7.14196 | 0.92971 | **0.02254791432** - filter bench: `bloom filter`: **-0.78% in ms/key** - ` ./filter_bench -impl=2 -quick -reserve_table_builder_memory=true | grep 'Build avg'` #-run | (pre-PR) avg ns/key | std ns/key | (post-PR) ns/key | std ns/key | change (%) -- | -- | -- | -- | -- | -- 10 | 26.4369 | 0.442182 | 26.3273 | 0.422919 | **-0.4145720565** 20 | 26.4451 | 0.592787 | 26.1419 | 0.62451 | **-1.1465262** - Crash test `python3 tools/db_crashtest.py blackbox --reserve_table_reader_memory=1 --cache_size=1` killed as normal Reviewed By: ajkr Differential Revision: D35136549 Pulled By: hx235 fbshipit-source-id: 146978858d0f900f43f4eb09bfd3e83195e3be28
3 years ago
#include <memory>
#include <unordered_set>
#include <vector>
#include "db/db_test_util.h"
#include "port/port.h"
#include "port/stack_trace.h"
#include "rocksdb/db.h"
Improve / clean up meta block code & integrity (#9163) Summary: * Checksums are now checked on meta blocks unless specifically suppressed or not applicable (e.g. plain table). (Was other way around.) This means a number of cases that were not checking checksums now are, including direct read TableProperties in Version::GetTableProperties (fixed in meta_blocks ReadTableProperties), reading any block from PersistentCache (fixed in BlockFetcher), read TableProperties in SstFileDumper (ldb/sst_dump/BackupEngine) before table reader open, maybe more. * For that to work, I moved the global_seqno+TableProperties checksum logic to the shared table/ code, because that is used by many utilies such as SstFileDumper. * Also for that to work, we have to know when we're dealing with a block that has a checksum (trailer), so added that capability to Footer based on magic number, and from there BlockFetcher. * Knowledge of trailer presence has also fixed a problem where other table formats were reading blocks including bytes for a non-existant trailer--and awkwardly kind-of not using them, e.g. no shared code checking checksums. (BlockFetcher compression type was populated incorrectly.) Now we only read what is needed. * Minimized code duplication and differing/incompatible/awkward abstractions in meta_blocks.{cc,h} (e.g. SeekTo in metaindex block without parsing block handle) * Moved some meta block handling code from table_properties*.* * Moved some code specific to block-based table from shared table/ code to BlockBasedTable class. The checksum stuff means we can't completely separate it, but things that don't need to be in shared table/ code should not be. * Use unique_ptr rather than raw ptr in more places. (Note: you can std::move from unique_ptr to shared_ptr.) Without enhancements to GetPropertiesOfAllTablesTest (see below), net reduction of roughly 100 lines of code. Pull Request resolved: https://github.com/facebook/rocksdb/pull/9163 Test Plan: existing tests and * Enhanced DBTablePropertiesTest.GetPropertiesOfAllTablesTest to verify that checksums are now checked on direct read of table properties by TableCache (new test would fail before this change) * Also enhanced DBTablePropertiesTest.GetPropertiesOfAllTablesTest to test putting table properties under old meta name * Also generally enhanced that same test to actually test what it was supposed to be testing already, by kicking things out of table cache when we don't want them there. Reviewed By: ajkr, mrambacher Differential Revision: D32514757 Pulled By: pdillinger fbshipit-source-id: 507964b9311d186ae8d1131182290cbd97a99fa9
3 years ago
#include "rocksdb/types.h"
#include "rocksdb/utilities/table_properties_collectors.h"
#include "table/format.h"
Improve / clean up meta block code & integrity (#9163) Summary: * Checksums are now checked on meta blocks unless specifically suppressed or not applicable (e.g. plain table). (Was other way around.) This means a number of cases that were not checking checksums now are, including direct read TableProperties in Version::GetTableProperties (fixed in meta_blocks ReadTableProperties), reading any block from PersistentCache (fixed in BlockFetcher), read TableProperties in SstFileDumper (ldb/sst_dump/BackupEngine) before table reader open, maybe more. * For that to work, I moved the global_seqno+TableProperties checksum logic to the shared table/ code, because that is used by many utilies such as SstFileDumper. * Also for that to work, we have to know when we're dealing with a block that has a checksum (trailer), so added that capability to Footer based on magic number, and from there BlockFetcher. * Knowledge of trailer presence has also fixed a problem where other table formats were reading blocks including bytes for a non-existant trailer--and awkwardly kind-of not using them, e.g. no shared code checking checksums. (BlockFetcher compression type was populated incorrectly.) Now we only read what is needed. * Minimized code duplication and differing/incompatible/awkward abstractions in meta_blocks.{cc,h} (e.g. SeekTo in metaindex block without parsing block handle) * Moved some meta block handling code from table_properties*.* * Moved some code specific to block-based table from shared table/ code to BlockBasedTable class. The checksum stuff means we can't completely separate it, but things that don't need to be in shared table/ code should not be. * Use unique_ptr rather than raw ptr in more places. (Note: you can std::move from unique_ptr to shared_ptr.) Without enhancements to GetPropertiesOfAllTablesTest (see below), net reduction of roughly 100 lines of code. Pull Request resolved: https://github.com/facebook/rocksdb/pull/9163 Test Plan: existing tests and * Enhanced DBTablePropertiesTest.GetPropertiesOfAllTablesTest to verify that checksums are now checked on direct read of table properties by TableCache (new test would fail before this change) * Also enhanced DBTablePropertiesTest.GetPropertiesOfAllTablesTest to test putting table properties under old meta name * Also generally enhanced that same test to actually test what it was supposed to be testing already, by kicking things out of table cache when we don't want them there. Reviewed By: ajkr, mrambacher Differential Revision: D32514757 Pulled By: pdillinger fbshipit-source-id: 507964b9311d186ae8d1131182290cbd97a99fa9
3 years ago
#include "table/meta_blocks.h"
Account memory of big memory users in BlockBasedTable in global memory limit (#9748) Summary: **Context:** Through heap profiling, we discovered that `BlockBasedTableReader` objects can accumulate and lead to high memory usage (e.g, `max_open_file = -1`). These memories are currently not saved, not tracked, not constrained and not cache evict-able. As a first step to improve this, similar to https://github.com/facebook/rocksdb/pull/8428, this PR is to track an estimate of `BlockBasedTableReader` object's memory in block cache and fail future creation if the memory usage exceeds the available space of cache at the time of creation. **Summary:** - Approximate big memory users (`BlockBasedTable::Rep` and `TableProperties` )' memory usage in addition to the existing estimated ones (filter block/index block/un-compression dictionary) - Charge all of these memory usages to block cache on `BlockBasedTable::Open()` and release them on `~BlockBasedTable()` as there is no memory usage fluctuation of concern in between - Refactor on CacheReservationManager (and its call-sites) to add concurrent support for BlockBasedTable used in this PR. Pull Request resolved: https://github.com/facebook/rocksdb/pull/9748 Test Plan: - New unit tests - db bench: `OpenDb` : **-0.52% in ms** - Setup `./db_bench -benchmarks=fillseq -db=/dev/shm/testdb -disable_auto_compactions=1 -write_buffer_size=1048576` - Repeated run with pre-change w/o feature and post-change with feature, benchmark `OpenDb`: `./db_bench -benchmarks=readrandom -use_existing_db=1 -db=/dev/shm/testdb -reserve_table_reader_memory=true (remove this when running w/o feature) -file_opening_threads=3 -open_files=-1 -report_open_timing=true| egrep 'OpenDb:'` #-run | (feature-off) avg milliseconds | std milliseconds | (feature-on) avg milliseconds | std milliseconds | change (%) -- | -- | -- | -- | -- | -- 10 | 11.4018 | 5.95173 | 9.47788 | 1.57538 | -16.87382694 20 | 9.23746 | 0.841053 | 9.32377 | 1.14074 | 0.9343477536 40 | 9.0876 | 0.671129 | 9.35053 | 1.11713 | 2.893283155 80 | 9.72514 | 2.28459 | 9.52013 | 1.0894 | -2.108041632 160 | 9.74677 | 0.991234 | 9.84743 | 1.73396 | 1.032752389 320 | 10.7297 | 5.11555 | 10.547 | 1.97692 | **-1.70275031** 640 | 11.7092 | 2.36565 | 11.7869 | 2.69377 | **0.6635807741** - db bench on write with cost to cache in WriteBufferManager (just in case this PR's CRM refactoring accidentally slows down anything in WBM) : `fillseq` : **+0.54% in micros/op** `./db_bench -benchmarks=fillseq -db=/dev/shm/testdb -disable_auto_compactions=1 -cost_write_buffer_to_cache=true -write_buffer_size=10000000000 | egrep 'fillseq'` #-run | (pre-PR) avg micros/op | std micros/op | (post-PR) avg micros/op | std micros/op | change (%) -- | -- | -- | -- | -- | -- 10 | 6.15 | 0.260187 | 6.289 | 0.371192 | 2.260162602 20 | 7.28025 | 0.465402 | 7.37255 | 0.451256 | 1.267813605 40 | 7.06312 | 0.490654 | 7.13803 | 0.478676 | **1.060579461** 80 | 7.14035 | 0.972831 | 7.14196 | 0.92971 | **0.02254791432** - filter bench: `bloom filter`: **-0.78% in ms/key** - ` ./filter_bench -impl=2 -quick -reserve_table_builder_memory=true | grep 'Build avg'` #-run | (pre-PR) avg ns/key | std ns/key | (post-PR) ns/key | std ns/key | change (%) -- | -- | -- | -- | -- | -- 10 | 26.4369 | 0.442182 | 26.3273 | 0.422919 | **-0.4145720565** 20 | 26.4451 | 0.592787 | 26.1419 | 0.62451 | **-1.1465262** - Crash test `python3 tools/db_crashtest.py blackbox --reserve_table_reader_memory=1 --cache_size=1` killed as normal Reviewed By: ajkr Differential Revision: D35136549 Pulled By: hx235 fbshipit-source-id: 146978858d0f900f43f4eb09bfd3e83195e3be28
3 years ago
#include "table/table_properties_internal.h"
#include "test_util/testharness.h"
#include "test_util/testutil.h"
#include "util/random.h"
namespace ROCKSDB_NAMESPACE {
// A helper function that ensures the table properties returned in
// `GetPropertiesOfAllTablesTest` is correct.
// This test assumes entries size is different for each of the tables.
namespace {
void VerifyTableProperties(DB* db, uint64_t expected_entries_size) {
TablePropertiesCollection props;
ASSERT_OK(db->GetPropertiesOfAllTables(&props));
ASSERT_EQ(4U, props.size());
std::unordered_set<uint64_t> unique_entries;
// Indirect test
uint64_t sum = 0;
for (const auto& item : props) {
unique_entries.insert(item.second->num_entries);
sum += item.second->num_entries;
}
ASSERT_EQ(props.size(), unique_entries.size());
ASSERT_EQ(expected_entries_size, sum);
Experimental support for SST unique IDs (#8990) Summary: * New public header unique_id.h and function GetUniqueIdFromTableProperties which computes a universally unique identifier based on table properties of table files from recent RocksDB versions. * Generation of DB session IDs is refactored so that they are guaranteed unique in the lifetime of a process running RocksDB. (SemiStructuredUniqueIdGen, new test included.) Along with file numbers, this enables SST unique IDs to be guaranteed unique among SSTs generated in a single process, and "better than random" between processes. See https://github.com/pdillinger/unique_id * In addition to public API producing 'external' unique IDs, there is a function for producing 'internal' unique IDs, with functions for converting between the two. In short, the external ID is "safe" for things people might do with it, and the internal ID enables more "power user" features for the future. Specifically, the external ID goes through a hashing layer so that any subset of bits in the external ID can be used as a hash of the full ID, while also preserving uniqueness guarantees in the first 128 bits (bijective both on first 128 bits and on full 192 bits). Intended follow-up: * Use the internal unique IDs in cache keys. (Avoid conflicts with https://github.com/facebook/rocksdb/issues/8912) (The file offset can be XORed into the third 64-bit value of the unique ID.) * Publish the external unique IDs in FileStorageInfo (https://github.com/facebook/rocksdb/issues/8968) Pull Request resolved: https://github.com/facebook/rocksdb/pull/8990 Test Plan: Unit tests added, and checking of unique ids in stress test. NOTE in stress test we do not generate nearly enough files to thoroughly stress uniqueness, but the test trims off pieces of the ID to check for uniqueness so that we can infer (with some assumptions) stronger properties in the aggregate. Reviewed By: zhichao-cao, mrambacher Differential Revision: D31582865 Pulled By: pdillinger fbshipit-source-id: 1f620c4c86af9abe2a8d177b9ccf2ad2b9f48243
3 years ago
VerifySstUniqueIds(props);
}
} // anonymous namespace
class DBTablePropertiesTest : public DBTestBase,
public testing::WithParamInterface<std::string> {
public:
DBTablePropertiesTest()
: DBTestBase("db_table_properties_test", /*env_do_fsync=*/false) {}
TablePropertiesCollection TestGetPropertiesOfTablesInRange(
std::vector<Range> ranges, std::size_t* num_properties = nullptr,
std::size_t* num_files = nullptr);
};
TEST_F(DBTablePropertiesTest, GetPropertiesOfAllTablesTest) {
Options options = CurrentOptions();
options.level0_file_num_compaction_trigger = 8;
Improve / clean up meta block code & integrity (#9163) Summary: * Checksums are now checked on meta blocks unless specifically suppressed or not applicable (e.g. plain table). (Was other way around.) This means a number of cases that were not checking checksums now are, including direct read TableProperties in Version::GetTableProperties (fixed in meta_blocks ReadTableProperties), reading any block from PersistentCache (fixed in BlockFetcher), read TableProperties in SstFileDumper (ldb/sst_dump/BackupEngine) before table reader open, maybe more. * For that to work, I moved the global_seqno+TableProperties checksum logic to the shared table/ code, because that is used by many utilies such as SstFileDumper. * Also for that to work, we have to know when we're dealing with a block that has a checksum (trailer), so added that capability to Footer based on magic number, and from there BlockFetcher. * Knowledge of trailer presence has also fixed a problem where other table formats were reading blocks including bytes for a non-existant trailer--and awkwardly kind-of not using them, e.g. no shared code checking checksums. (BlockFetcher compression type was populated incorrectly.) Now we only read what is needed. * Minimized code duplication and differing/incompatible/awkward abstractions in meta_blocks.{cc,h} (e.g. SeekTo in metaindex block without parsing block handle) * Moved some meta block handling code from table_properties*.* * Moved some code specific to block-based table from shared table/ code to BlockBasedTable class. The checksum stuff means we can't completely separate it, but things that don't need to be in shared table/ code should not be. * Use unique_ptr rather than raw ptr in more places. (Note: you can std::move from unique_ptr to shared_ptr.) Without enhancements to GetPropertiesOfAllTablesTest (see below), net reduction of roughly 100 lines of code. Pull Request resolved: https://github.com/facebook/rocksdb/pull/9163 Test Plan: existing tests and * Enhanced DBTablePropertiesTest.GetPropertiesOfAllTablesTest to verify that checksums are now checked on direct read of table properties by TableCache (new test would fail before this change) * Also enhanced DBTablePropertiesTest.GetPropertiesOfAllTablesTest to test putting table properties under old meta name * Also generally enhanced that same test to actually test what it was supposed to be testing already, by kicking things out of table cache when we don't want them there. Reviewed By: ajkr, mrambacher Differential Revision: D32514757 Pulled By: pdillinger fbshipit-source-id: 507964b9311d186ae8d1131182290cbd97a99fa9
3 years ago
// Part of strategy to prevent pinning table files
options.max_open_files = 42;
Reopen(options);
Improve / clean up meta block code & integrity (#9163) Summary: * Checksums are now checked on meta blocks unless specifically suppressed or not applicable (e.g. plain table). (Was other way around.) This means a number of cases that were not checking checksums now are, including direct read TableProperties in Version::GetTableProperties (fixed in meta_blocks ReadTableProperties), reading any block from PersistentCache (fixed in BlockFetcher), read TableProperties in SstFileDumper (ldb/sst_dump/BackupEngine) before table reader open, maybe more. * For that to work, I moved the global_seqno+TableProperties checksum logic to the shared table/ code, because that is used by many utilies such as SstFileDumper. * Also for that to work, we have to know when we're dealing with a block that has a checksum (trailer), so added that capability to Footer based on magic number, and from there BlockFetcher. * Knowledge of trailer presence has also fixed a problem where other table formats were reading blocks including bytes for a non-existant trailer--and awkwardly kind-of not using them, e.g. no shared code checking checksums. (BlockFetcher compression type was populated incorrectly.) Now we only read what is needed. * Minimized code duplication and differing/incompatible/awkward abstractions in meta_blocks.{cc,h} (e.g. SeekTo in metaindex block without parsing block handle) * Moved some meta block handling code from table_properties*.* * Moved some code specific to block-based table from shared table/ code to BlockBasedTable class. The checksum stuff means we can't completely separate it, but things that don't need to be in shared table/ code should not be. * Use unique_ptr rather than raw ptr in more places. (Note: you can std::move from unique_ptr to shared_ptr.) Without enhancements to GetPropertiesOfAllTablesTest (see below), net reduction of roughly 100 lines of code. Pull Request resolved: https://github.com/facebook/rocksdb/pull/9163 Test Plan: existing tests and * Enhanced DBTablePropertiesTest.GetPropertiesOfAllTablesTest to verify that checksums are now checked on direct read of table properties by TableCache (new test would fail before this change) * Also enhanced DBTablePropertiesTest.GetPropertiesOfAllTablesTest to test putting table properties under old meta name * Also generally enhanced that same test to actually test what it was supposed to be testing already, by kicking things out of table cache when we don't want them there. Reviewed By: ajkr, mrambacher Differential Revision: D32514757 Pulled By: pdillinger fbshipit-source-id: 507964b9311d186ae8d1131182290cbd97a99fa9
3 years ago
// Create 4 tables
for (int table = 0; table < 4; ++table) {
Improve / clean up meta block code & integrity (#9163) Summary: * Checksums are now checked on meta blocks unless specifically suppressed or not applicable (e.g. plain table). (Was other way around.) This means a number of cases that were not checking checksums now are, including direct read TableProperties in Version::GetTableProperties (fixed in meta_blocks ReadTableProperties), reading any block from PersistentCache (fixed in BlockFetcher), read TableProperties in SstFileDumper (ldb/sst_dump/BackupEngine) before table reader open, maybe more. * For that to work, I moved the global_seqno+TableProperties checksum logic to the shared table/ code, because that is used by many utilies such as SstFileDumper. * Also for that to work, we have to know when we're dealing with a block that has a checksum (trailer), so added that capability to Footer based on magic number, and from there BlockFetcher. * Knowledge of trailer presence has also fixed a problem where other table formats were reading blocks including bytes for a non-existant trailer--and awkwardly kind-of not using them, e.g. no shared code checking checksums. (BlockFetcher compression type was populated incorrectly.) Now we only read what is needed. * Minimized code duplication and differing/incompatible/awkward abstractions in meta_blocks.{cc,h} (e.g. SeekTo in metaindex block without parsing block handle) * Moved some meta block handling code from table_properties*.* * Moved some code specific to block-based table from shared table/ code to BlockBasedTable class. The checksum stuff means we can't completely separate it, but things that don't need to be in shared table/ code should not be. * Use unique_ptr rather than raw ptr in more places. (Note: you can std::move from unique_ptr to shared_ptr.) Without enhancements to GetPropertiesOfAllTablesTest (see below), net reduction of roughly 100 lines of code. Pull Request resolved: https://github.com/facebook/rocksdb/pull/9163 Test Plan: existing tests and * Enhanced DBTablePropertiesTest.GetPropertiesOfAllTablesTest to verify that checksums are now checked on direct read of table properties by TableCache (new test would fail before this change) * Also enhanced DBTablePropertiesTest.GetPropertiesOfAllTablesTest to test putting table properties under old meta name * Also generally enhanced that same test to actually test what it was supposed to be testing already, by kicking things out of table cache when we don't want them there. Reviewed By: ajkr, mrambacher Differential Revision: D32514757 Pulled By: pdillinger fbshipit-source-id: 507964b9311d186ae8d1131182290cbd97a99fa9
3 years ago
// Use old meta name for table properties for one file
if (table == 3) {
SyncPoint::GetInstance()->SetCallBack(
"BlockBasedTableBuilder::WritePropertiesBlock:Meta", [&](void* meta) {
*reinterpret_cast<const std::string**>(meta) =
&kPropertiesBlockOldName;
});
SyncPoint::GetInstance()->EnableProcessing();
}
// Build file
for (int i = 0; i < 10 + table; ++i) {
ASSERT_OK(
db_->Put(WriteOptions(), std::to_string(table * 100 + i), "val"));
}
ASSERT_OK(db_->Flush(FlushOptions()));
}
Improve / clean up meta block code & integrity (#9163) Summary: * Checksums are now checked on meta blocks unless specifically suppressed or not applicable (e.g. plain table). (Was other way around.) This means a number of cases that were not checking checksums now are, including direct read TableProperties in Version::GetTableProperties (fixed in meta_blocks ReadTableProperties), reading any block from PersistentCache (fixed in BlockFetcher), read TableProperties in SstFileDumper (ldb/sst_dump/BackupEngine) before table reader open, maybe more. * For that to work, I moved the global_seqno+TableProperties checksum logic to the shared table/ code, because that is used by many utilies such as SstFileDumper. * Also for that to work, we have to know when we're dealing with a block that has a checksum (trailer), so added that capability to Footer based on magic number, and from there BlockFetcher. * Knowledge of trailer presence has also fixed a problem where other table formats were reading blocks including bytes for a non-existant trailer--and awkwardly kind-of not using them, e.g. no shared code checking checksums. (BlockFetcher compression type was populated incorrectly.) Now we only read what is needed. * Minimized code duplication and differing/incompatible/awkward abstractions in meta_blocks.{cc,h} (e.g. SeekTo in metaindex block without parsing block handle) * Moved some meta block handling code from table_properties*.* * Moved some code specific to block-based table from shared table/ code to BlockBasedTable class. The checksum stuff means we can't completely separate it, but things that don't need to be in shared table/ code should not be. * Use unique_ptr rather than raw ptr in more places. (Note: you can std::move from unique_ptr to shared_ptr.) Without enhancements to GetPropertiesOfAllTablesTest (see below), net reduction of roughly 100 lines of code. Pull Request resolved: https://github.com/facebook/rocksdb/pull/9163 Test Plan: existing tests and * Enhanced DBTablePropertiesTest.GetPropertiesOfAllTablesTest to verify that checksums are now checked on direct read of table properties by TableCache (new test would fail before this change) * Also enhanced DBTablePropertiesTest.GetPropertiesOfAllTablesTest to test putting table properties under old meta name * Also generally enhanced that same test to actually test what it was supposed to be testing already, by kicking things out of table cache when we don't want them there. Reviewed By: ajkr, mrambacher Differential Revision: D32514757 Pulled By: pdillinger fbshipit-source-id: 507964b9311d186ae8d1131182290cbd97a99fa9
3 years ago
SyncPoint::GetInstance()->DisableProcessing();
std::string original_session_id;
ASSERT_OK(db_->GetDbSessionId(original_session_id));
// Part of strategy to prevent pinning table files
SyncPoint::GetInstance()->SetCallBack(
"VersionEditHandler::LoadTables:skip_load_table_files",
[&](void* skip_load) { *reinterpret_cast<bool*>(skip_load) = true; });
SyncPoint::GetInstance()->EnableProcessing();
// 1. Read table properties directly from file
Reopen(options);
Improve / clean up meta block code & integrity (#9163) Summary: * Checksums are now checked on meta blocks unless specifically suppressed or not applicable (e.g. plain table). (Was other way around.) This means a number of cases that were not checking checksums now are, including direct read TableProperties in Version::GetTableProperties (fixed in meta_blocks ReadTableProperties), reading any block from PersistentCache (fixed in BlockFetcher), read TableProperties in SstFileDumper (ldb/sst_dump/BackupEngine) before table reader open, maybe more. * For that to work, I moved the global_seqno+TableProperties checksum logic to the shared table/ code, because that is used by many utilies such as SstFileDumper. * Also for that to work, we have to know when we're dealing with a block that has a checksum (trailer), so added that capability to Footer based on magic number, and from there BlockFetcher. * Knowledge of trailer presence has also fixed a problem where other table formats were reading blocks including bytes for a non-existant trailer--and awkwardly kind-of not using them, e.g. no shared code checking checksums. (BlockFetcher compression type was populated incorrectly.) Now we only read what is needed. * Minimized code duplication and differing/incompatible/awkward abstractions in meta_blocks.{cc,h} (e.g. SeekTo in metaindex block without parsing block handle) * Moved some meta block handling code from table_properties*.* * Moved some code specific to block-based table from shared table/ code to BlockBasedTable class. The checksum stuff means we can't completely separate it, but things that don't need to be in shared table/ code should not be. * Use unique_ptr rather than raw ptr in more places. (Note: you can std::move from unique_ptr to shared_ptr.) Without enhancements to GetPropertiesOfAllTablesTest (see below), net reduction of roughly 100 lines of code. Pull Request resolved: https://github.com/facebook/rocksdb/pull/9163 Test Plan: existing tests and * Enhanced DBTablePropertiesTest.GetPropertiesOfAllTablesTest to verify that checksums are now checked on direct read of table properties by TableCache (new test would fail before this change) * Also enhanced DBTablePropertiesTest.GetPropertiesOfAllTablesTest to test putting table properties under old meta name * Also generally enhanced that same test to actually test what it was supposed to be testing already, by kicking things out of table cache when we don't want them there. Reviewed By: ajkr, mrambacher Differential Revision: D32514757 Pulled By: pdillinger fbshipit-source-id: 507964b9311d186ae8d1131182290cbd97a99fa9
3 years ago
// Clear out auto-opened files
dbfull()->TEST_table_cache()->EraseUnRefEntries();
ASSERT_EQ(dbfull()->TEST_table_cache()->GetUsage(), 0U);
VerifyTableProperties(db_, 10 + 11 + 12 + 13);
// 2. Put two tables to table cache and
Reopen(options);
Improve / clean up meta block code & integrity (#9163) Summary: * Checksums are now checked on meta blocks unless specifically suppressed or not applicable (e.g. plain table). (Was other way around.) This means a number of cases that were not checking checksums now are, including direct read TableProperties in Version::GetTableProperties (fixed in meta_blocks ReadTableProperties), reading any block from PersistentCache (fixed in BlockFetcher), read TableProperties in SstFileDumper (ldb/sst_dump/BackupEngine) before table reader open, maybe more. * For that to work, I moved the global_seqno+TableProperties checksum logic to the shared table/ code, because that is used by many utilies such as SstFileDumper. * Also for that to work, we have to know when we're dealing with a block that has a checksum (trailer), so added that capability to Footer based on magic number, and from there BlockFetcher. * Knowledge of trailer presence has also fixed a problem where other table formats were reading blocks including bytes for a non-existant trailer--and awkwardly kind-of not using them, e.g. no shared code checking checksums. (BlockFetcher compression type was populated incorrectly.) Now we only read what is needed. * Minimized code duplication and differing/incompatible/awkward abstractions in meta_blocks.{cc,h} (e.g. SeekTo in metaindex block without parsing block handle) * Moved some meta block handling code from table_properties*.* * Moved some code specific to block-based table from shared table/ code to BlockBasedTable class. The checksum stuff means we can't completely separate it, but things that don't need to be in shared table/ code should not be. * Use unique_ptr rather than raw ptr in more places. (Note: you can std::move from unique_ptr to shared_ptr.) Without enhancements to GetPropertiesOfAllTablesTest (see below), net reduction of roughly 100 lines of code. Pull Request resolved: https://github.com/facebook/rocksdb/pull/9163 Test Plan: existing tests and * Enhanced DBTablePropertiesTest.GetPropertiesOfAllTablesTest to verify that checksums are now checked on direct read of table properties by TableCache (new test would fail before this change) * Also enhanced DBTablePropertiesTest.GetPropertiesOfAllTablesTest to test putting table properties under old meta name * Also generally enhanced that same test to actually test what it was supposed to be testing already, by kicking things out of table cache when we don't want them there. Reviewed By: ajkr, mrambacher Differential Revision: D32514757 Pulled By: pdillinger fbshipit-source-id: 507964b9311d186ae8d1131182290cbd97a99fa9
3 years ago
// Clear out auto-opened files
dbfull()->TEST_table_cache()->EraseUnRefEntries();
ASSERT_EQ(dbfull()->TEST_table_cache()->GetUsage(), 0U);
// fetch key from 1st and 2nd table, which will internally place that table to
// the table cache.
for (int i = 0; i < 2; ++i) {
Get(std::to_string(i * 100 + 0));
}
VerifyTableProperties(db_, 10 + 11 + 12 + 13);
// 3. Put all tables to table cache
Reopen(options);
Improve / clean up meta block code & integrity (#9163) Summary: * Checksums are now checked on meta blocks unless specifically suppressed or not applicable (e.g. plain table). (Was other way around.) This means a number of cases that were not checking checksums now are, including direct read TableProperties in Version::GetTableProperties (fixed in meta_blocks ReadTableProperties), reading any block from PersistentCache (fixed in BlockFetcher), read TableProperties in SstFileDumper (ldb/sst_dump/BackupEngine) before table reader open, maybe more. * For that to work, I moved the global_seqno+TableProperties checksum logic to the shared table/ code, because that is used by many utilies such as SstFileDumper. * Also for that to work, we have to know when we're dealing with a block that has a checksum (trailer), so added that capability to Footer based on magic number, and from there BlockFetcher. * Knowledge of trailer presence has also fixed a problem where other table formats were reading blocks including bytes for a non-existant trailer--and awkwardly kind-of not using them, e.g. no shared code checking checksums. (BlockFetcher compression type was populated incorrectly.) Now we only read what is needed. * Minimized code duplication and differing/incompatible/awkward abstractions in meta_blocks.{cc,h} (e.g. SeekTo in metaindex block without parsing block handle) * Moved some meta block handling code from table_properties*.* * Moved some code specific to block-based table from shared table/ code to BlockBasedTable class. The checksum stuff means we can't completely separate it, but things that don't need to be in shared table/ code should not be. * Use unique_ptr rather than raw ptr in more places. (Note: you can std::move from unique_ptr to shared_ptr.) Without enhancements to GetPropertiesOfAllTablesTest (see below), net reduction of roughly 100 lines of code. Pull Request resolved: https://github.com/facebook/rocksdb/pull/9163 Test Plan: existing tests and * Enhanced DBTablePropertiesTest.GetPropertiesOfAllTablesTest to verify that checksums are now checked on direct read of table properties by TableCache (new test would fail before this change) * Also enhanced DBTablePropertiesTest.GetPropertiesOfAllTablesTest to test putting table properties under old meta name * Also generally enhanced that same test to actually test what it was supposed to be testing already, by kicking things out of table cache when we don't want them there. Reviewed By: ajkr, mrambacher Differential Revision: D32514757 Pulled By: pdillinger fbshipit-source-id: 507964b9311d186ae8d1131182290cbd97a99fa9
3 years ago
// fetch key from all tables, which will place them in table cache.
for (int i = 0; i < 4; ++i) {
Get(std::to_string(i * 100 + 0));
}
VerifyTableProperties(db_, 10 + 11 + 12 + 13);
Improve / clean up meta block code & integrity (#9163) Summary: * Checksums are now checked on meta blocks unless specifically suppressed or not applicable (e.g. plain table). (Was other way around.) This means a number of cases that were not checking checksums now are, including direct read TableProperties in Version::GetTableProperties (fixed in meta_blocks ReadTableProperties), reading any block from PersistentCache (fixed in BlockFetcher), read TableProperties in SstFileDumper (ldb/sst_dump/BackupEngine) before table reader open, maybe more. * For that to work, I moved the global_seqno+TableProperties checksum logic to the shared table/ code, because that is used by many utilies such as SstFileDumper. * Also for that to work, we have to know when we're dealing with a block that has a checksum (trailer), so added that capability to Footer based on magic number, and from there BlockFetcher. * Knowledge of trailer presence has also fixed a problem where other table formats were reading blocks including bytes for a non-existant trailer--and awkwardly kind-of not using them, e.g. no shared code checking checksums. (BlockFetcher compression type was populated incorrectly.) Now we only read what is needed. * Minimized code duplication and differing/incompatible/awkward abstractions in meta_blocks.{cc,h} (e.g. SeekTo in metaindex block without parsing block handle) * Moved some meta block handling code from table_properties*.* * Moved some code specific to block-based table from shared table/ code to BlockBasedTable class. The checksum stuff means we can't completely separate it, but things that don't need to be in shared table/ code should not be. * Use unique_ptr rather than raw ptr in more places. (Note: you can std::move from unique_ptr to shared_ptr.) Without enhancements to GetPropertiesOfAllTablesTest (see below), net reduction of roughly 100 lines of code. Pull Request resolved: https://github.com/facebook/rocksdb/pull/9163 Test Plan: existing tests and * Enhanced DBTablePropertiesTest.GetPropertiesOfAllTablesTest to verify that checksums are now checked on direct read of table properties by TableCache (new test would fail before this change) * Also enhanced DBTablePropertiesTest.GetPropertiesOfAllTablesTest to test putting table properties under old meta name * Also generally enhanced that same test to actually test what it was supposed to be testing already, by kicking things out of table cache when we don't want them there. Reviewed By: ajkr, mrambacher Differential Revision: D32514757 Pulled By: pdillinger fbshipit-source-id: 507964b9311d186ae8d1131182290cbd97a99fa9
3 years ago
// 4. Try to read CORRUPT properties (a) directly from file, and (b)
// through reader on Get
// It's not practical to prevent table file read on Open, so we
// corrupt after open and after purging table cache.
for (bool direct : {true, false}) {
Reopen(options);
// Clear out auto-opened files
dbfull()->TEST_table_cache()->EraseUnRefEntries();
ASSERT_EQ(dbfull()->TEST_table_cache()->GetUsage(), 0U);
TablePropertiesCollection props;
ASSERT_OK(db_->GetPropertiesOfAllTables(&props));
std::string sst_file = props.begin()->first;
// Corrupt the file's TableProperties using session id
std::string contents;
ASSERT_OK(
ReadFileToString(env_->GetFileSystem().get(), sst_file, &contents));
size_t pos = contents.find(original_session_id);
ASSERT_NE(pos, std::string::npos);
ASSERT_OK(test::CorruptFile(env_, sst_file, static_cast<int>(pos), 1,
/*verify checksum fails*/ false));
// Try to read CORRUPT properties
if (direct) {
ASSERT_TRUE(db_->GetPropertiesOfAllTables(&props).IsCorruption());
} else {
bool found_corruption = false;
for (int i = 0; i < 4; ++i) {
std::string result = Get(std::to_string(i * 100 + 0));
Improve / clean up meta block code & integrity (#9163) Summary: * Checksums are now checked on meta blocks unless specifically suppressed or not applicable (e.g. plain table). (Was other way around.) This means a number of cases that were not checking checksums now are, including direct read TableProperties in Version::GetTableProperties (fixed in meta_blocks ReadTableProperties), reading any block from PersistentCache (fixed in BlockFetcher), read TableProperties in SstFileDumper (ldb/sst_dump/BackupEngine) before table reader open, maybe more. * For that to work, I moved the global_seqno+TableProperties checksum logic to the shared table/ code, because that is used by many utilies such as SstFileDumper. * Also for that to work, we have to know when we're dealing with a block that has a checksum (trailer), so added that capability to Footer based on magic number, and from there BlockFetcher. * Knowledge of trailer presence has also fixed a problem where other table formats were reading blocks including bytes for a non-existant trailer--and awkwardly kind-of not using them, e.g. no shared code checking checksums. (BlockFetcher compression type was populated incorrectly.) Now we only read what is needed. * Minimized code duplication and differing/incompatible/awkward abstractions in meta_blocks.{cc,h} (e.g. SeekTo in metaindex block without parsing block handle) * Moved some meta block handling code from table_properties*.* * Moved some code specific to block-based table from shared table/ code to BlockBasedTable class. The checksum stuff means we can't completely separate it, but things that don't need to be in shared table/ code should not be. * Use unique_ptr rather than raw ptr in more places. (Note: you can std::move from unique_ptr to shared_ptr.) Without enhancements to GetPropertiesOfAllTablesTest (see below), net reduction of roughly 100 lines of code. Pull Request resolved: https://github.com/facebook/rocksdb/pull/9163 Test Plan: existing tests and * Enhanced DBTablePropertiesTest.GetPropertiesOfAllTablesTest to verify that checksums are now checked on direct read of table properties by TableCache (new test would fail before this change) * Also enhanced DBTablePropertiesTest.GetPropertiesOfAllTablesTest to test putting table properties under old meta name * Also generally enhanced that same test to actually test what it was supposed to be testing already, by kicking things out of table cache when we don't want them there. Reviewed By: ajkr, mrambacher Differential Revision: D32514757 Pulled By: pdillinger fbshipit-source-id: 507964b9311d186ae8d1131182290cbd97a99fa9
3 years ago
if (result.find_first_of("Corruption: block checksum mismatch") !=
std::string::npos) {
found_corruption = true;
}
}
ASSERT_TRUE(found_corruption);
}
// UN-corrupt file for next iteration
ASSERT_OK(test::CorruptFile(env_, sst_file, static_cast<int>(pos), 1,
/*verify checksum fails*/ false));
}
SyncPoint::GetInstance()->DisableProcessing();
}
TEST_F(DBTablePropertiesTest, InvalidIgnored) {
// RocksDB versions 2.5 - 2.7 generate some properties that Block considers
// invalid in some way. This approximates that.
// Inject properties block data that Block considers invalid
SyncPoint::GetInstance()->SetCallBack(
"BlockBasedTableBuilder::WritePropertiesBlock:BlockData",
[&](void* block_data) {
*reinterpret_cast<Slice*>(block_data) = Slice("X");
});
SyncPoint::GetInstance()->EnableProcessing();
Always verify SST unique IDs on SST file open (#10532) Summary: Although we've been tracking SST unique IDs in the DB manifest unconditionally, checking has been opt-in and with an extra pass at DB::Open time. This changes the behavior of `verify_sst_unique_id_in_manifest` to check unique ID against manifest every time an SST file is opened through table cache (normal DB operations), replacing the explicit pass over files at DB::Open time. This change also enables the option by default and removes the "EXPERIMENTAL" designation. One possible criticism is that the option no longer ensures the integrity of a DB at Open time. This is far from an all-or-nothing issue. Verifying the IDs of all SST files hardly ensures all the data in the DB is readable. (VerifyChecksum is supposed to do that.) Also, with max_open_files=-1 (default, extremely common), all SST files are opened at DB::Open time anyway. Implementation details: * `VerifySstUniqueIdInManifest()` functions are the extra/explicit pass that is now removed. * Unit tests that manipulate/corrupt table properties have to opt out of this check, because that corrupts the "actual" unique id. (And even for testing we don't currently have a mechanism to set "no unique id" in the in-memory file metadata for new files.) * A lot of other unit test churn relates to (a) default checking on, and (b) checking on SST open even without DB::Open (e.g. on flush) * Use `FileMetaData` for more `TableCache` operations (in place of `FileDescriptor`) so that we have access to the unique_id whenever we might need to open an SST file. **There is the possibility of performance impact because we can no longer use the more localized `fd` part of an `FdWithKeyRange` but instead follow the `file_metadata` pointer. However, this change (possible regression) is only done for `GetMemoryUsageByTableReaders`.** * Removed a completely unnecessary constructor overload of `TableReaderOptions` Possible follow-up: * Verification only happens when opening through table cache. Are there more places where this should happen? * Improve error message when there is a file size mismatch vs. manifest (FIXME added in the appropriate place). * I'm not sure there's a justification for `FileDescriptor` to be distinct from `FileMetaData`. * I'm skeptical that `FdWithKeyRange` really still makes sense for optimizing some data locality by duplicating some data in memory, but I could be wrong. * An unnecessary overload of NewTableReader was recently added, in the public API nonetheless (though unusable there). It should be cleaned up to put most things under `TableReaderOptions`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/10532 Test Plan: updated unit tests Performance test showing no significant difference (just noise I think): `./db_bench -benchmarks=readwhilewriting[-X10] -num=3000000 -disable_wal=1 -bloom_bits=8 -write_buffer_size=1000000 -target_file_size_base=1000000` Before: readwhilewriting [AVG 10 runs] : 68702 (± 6932) ops/sec After: readwhilewriting [AVG 10 runs] : 68239 (± 7198) ops/sec Reviewed By: jay-zhuang Differential Revision: D38765551 Pulled By: pdillinger fbshipit-source-id: a827a708155f12344ab2a5c16e7701c7636da4c2
2 years ago
// Corrupting the table properties corrupts the unique id.
// Ignore the unique id recorded in the manifest.
auto options = CurrentOptions();
options.verify_sst_unique_id_in_manifest = false;
Reopen(options);
// Build file
for (int i = 0; i < 10; ++i) {
ASSERT_OK(db_->Put(WriteOptions(), std::to_string(i), "val"));
}
ASSERT_OK(db_->Flush(FlushOptions()));
SyncPoint::GetInstance()->DisableProcessing();
// Not crashing is good enough
TablePropertiesCollection props;
ASSERT_OK(db_->GetPropertiesOfAllTables(&props));
}
TEST_F(DBTablePropertiesTest, CreateOnDeletionCollectorFactory) {
ConfigOptions options;
options.ignore_unsupported_options = false;
std::shared_ptr<TablePropertiesCollectorFactory> factory;
std::string id = CompactOnDeletionCollectorFactory::kClassName();
ASSERT_OK(
TablePropertiesCollectorFactory::CreateFromString(options, id, &factory));
auto del_factory = factory->CheckedCast<CompactOnDeletionCollectorFactory>();
ASSERT_NE(del_factory, nullptr);
ASSERT_EQ(0U, del_factory->GetWindowSize());
ASSERT_EQ(0U, del_factory->GetDeletionTrigger());
ASSERT_EQ(0.0, del_factory->GetDeletionRatio());
ASSERT_OK(TablePropertiesCollectorFactory::CreateFromString(
options, "window_size=100; deletion_trigger=90; id=" + id, &factory));
del_factory = factory->CheckedCast<CompactOnDeletionCollectorFactory>();
ASSERT_NE(del_factory, nullptr);
ASSERT_EQ(100U, del_factory->GetWindowSize());
ASSERT_EQ(90U, del_factory->GetDeletionTrigger());
ASSERT_EQ(0.0, del_factory->GetDeletionRatio());
ASSERT_OK(TablePropertiesCollectorFactory::CreateFromString(
options,
"window_size=100; deletion_trigger=90; deletion_ratio=0.5; id=" + id,
&factory));
del_factory = factory->CheckedCast<CompactOnDeletionCollectorFactory>();
ASSERT_NE(del_factory, nullptr);
ASSERT_EQ(100U, del_factory->GetWindowSize());
ASSERT_EQ(90U, del_factory->GetDeletionTrigger());
ASSERT_EQ(0.5, del_factory->GetDeletionRatio());
}
TablePropertiesCollection
DBTablePropertiesTest::TestGetPropertiesOfTablesInRange(
std::vector<Range> ranges, std::size_t* num_properties,
std::size_t* num_files) {
// Since we deref zero element in the vector it can not be empty
// otherwise we pass an address to some random memory
EXPECT_GT(ranges.size(), 0U);
// run the query
TablePropertiesCollection props;
EXPECT_OK(db_->GetPropertiesOfTablesInRange(
db_->DefaultColumnFamily(), &ranges[0], ranges.size(), &props));
// Make sure that we've received properties for those and for those files
// only which fall within requested ranges
std::vector<LiveFileMetaData> vmd;
db_->GetLiveFilesMetaData(&vmd);
for (auto& md : vmd) {
std::string fn = md.db_path + md.name;
bool in_range = false;
for (auto& r : ranges) {
// smallestkey < limit && largestkey >= start
if (r.limit.compare(md.smallestkey) >= 0 &&
r.start.compare(md.largestkey) <= 0) {
in_range = true;
EXPECT_GT(props.count(fn), 0);
}
}
if (!in_range) {
EXPECT_EQ(props.count(fn), 0);
}
}
if (num_properties) {
*num_properties = props.size();
}
if (num_files) {
*num_files = vmd.size();
}
return props;
}
TEST_F(DBTablePropertiesTest, GetPropertiesOfTablesInRange) {
// Fixed random sead
Random rnd(301);
Options options;
Make option `level_compaction_dynamic_level_bytes` true by default (#11525) Summary: after https://github.com/facebook/rocksdb/issues/11321 and https://github.com/facebook/rocksdb/issues/11340 (both included in RocksDB v8.2), migration from `level_compaction_dynamic_level_bytes=false` to `level_compaction_dynamic_level_bytes=true` is automatic by RocksDB and requires no manual compaction from user. Making the option true by default as it has several advantages: 1. better space amplification guarantee (a more stable LSM shape). 2. compaction is more adaptive to write traffic. 3. automatic draining of unneeded levels. Wiki is updated with more detail: https://github.com/facebook/rocksdb/wiki/Leveled-Compaction#option-level_compaction_dynamic_level_bytes-and-levels-target-size. The PR mostly contains fixes for unit tests as they assumed `level_compaction_dynamic_level_bytes=false`. Most notable change is commit https://github.com/facebook/rocksdb/pull/11525/commits/f742be330ca1a7abc33107b00df99818f71c387b and https://github.com/facebook/rocksdb/pull/11525/commits/b1928e42b34c0d4f1cc5d5239149870c6dc7a737 which override the default option in DBTestBase to still set `level_compaction_dynamic_level_bytes=false` by default. This helps to reduce the change needed for unit tests. I think this default option override in unit tests is okay since the behavior of `level_compaction_dynamic_level_bytes=true` is tested by explicitly setting this option. Also, `level_compaction_dynamic_level_bytes=false` may be more desired in unit tests as it makes it easier to create a desired LSM shape. Comment for option `level_compaction_dynamic_level_bytes` is updated to reflect this change and change made in https://github.com/facebook/rocksdb/issues/10057. Pull Request resolved: https://github.com/facebook/rocksdb/pull/11525 Test Plan: `make -j32 J=32 check` several times to try to catch flaky tests due to this option change. Reviewed By: ajkr Differential Revision: D46654256 Pulled By: cbi42 fbshipit-source-id: 6b5827dae124f6f1fdc8cca2ac6f6fcd878830e1
2 years ago
options.level_compaction_dynamic_level_bytes = false;
options.create_if_missing = true;
options.write_buffer_size = 4096;
options.max_write_buffer_number = 2;
options.level0_file_num_compaction_trigger = 2;
options.level0_slowdown_writes_trigger = 2;
options.level0_stop_writes_trigger = 2;
options.target_file_size_base = 2048;
options.max_bytes_for_level_base = 40960;
options.max_bytes_for_level_multiplier = 4;
options.hard_pending_compaction_bytes_limit = 16 * 1024;
options.num_levels = 8;
options.env = env_;
DestroyAndReopen(options);
// build a decent LSM
for (int i = 0; i < 10000; i++) {
ASSERT_OK(Put(test::RandomKey(&rnd, 5), rnd.RandomString(102)));
}
ASSERT_OK(Flush());
ASSERT_OK(dbfull()->TEST_WaitForCompact());
if (NumTableFilesAtLevel(0) == 0) {
ASSERT_OK(Put(test::RandomKey(&rnd, 5), rnd.RandomString(102)));
ASSERT_OK(Flush());
}
ASSERT_OK(db_->PauseBackgroundWork());
// Ensure that we have at least L0, L1 and L2
ASSERT_GT(NumTableFilesAtLevel(0), 0);
ASSERT_GT(NumTableFilesAtLevel(1), 0);
ASSERT_GT(NumTableFilesAtLevel(2), 0);
// Query the largest range
std::size_t num_properties, num_files;
TestGetPropertiesOfTablesInRange(
{Range(test::RandomKey(&rnd, 5, test::RandomKeyType::SMALLEST),
test::RandomKey(&rnd, 5, test::RandomKeyType::LARGEST))},
&num_properties, &num_files);
ASSERT_EQ(num_properties, num_files);
// Query the empty range
TestGetPropertiesOfTablesInRange(
{Range(test::RandomKey(&rnd, 5, test::RandomKeyType::LARGEST),
test::RandomKey(&rnd, 5, test::RandomKeyType::SMALLEST))},
&num_properties, &num_files);
ASSERT_GT(num_files, 0);
ASSERT_EQ(num_properties, 0);
// Query the middle rangee
TestGetPropertiesOfTablesInRange(
{Range(test::RandomKey(&rnd, 5, test::RandomKeyType::MIDDLE),
test::RandomKey(&rnd, 5, test::RandomKeyType::LARGEST))},
&num_properties, &num_files);
ASSERT_GT(num_files, 0);
ASSERT_GT(num_files, num_properties);
ASSERT_GT(num_properties, 0);
// Query a bunch of random ranges
for (int j = 0; j < 100; j++) {
// create a bunch of ranges
std::vector<std::string> random_keys;
// Random returns numbers with zero included
// when we pass empty ranges TestGetPropertiesOfTablesInRange()
// derefs random memory in the empty ranges[0]
// so want to be greater than zero and even since
// the below loop requires that random_keys.size() to be even.
auto n = 2 * (rnd.Uniform(50) + 1);
for (uint32_t i = 0; i < n; ++i) {
random_keys.push_back(test::RandomKey(&rnd, 5));
}
ASSERT_GT(random_keys.size(), 0U);
ASSERT_EQ((random_keys.size() % 2), 0U);
std::vector<Range> ranges;
auto it = random_keys.begin();
while (it != random_keys.end()) {
ranges.push_back(Range(*it, *(it + 1)));
it += 2;
}
TestGetPropertiesOfTablesInRange(std::move(ranges));
}
}
TEST_F(DBTablePropertiesTest, GetColumnFamilyNameProperty) {
std::string kExtraCfName = "pikachu";
CreateAndReopenWithCF({kExtraCfName}, CurrentOptions());
// Create one table per CF, then verify it was created with the column family
// name property.
for (uint32_t cf = 0; cf < 2; ++cf) {
ASSERT_OK(Put(cf, "key", "val"));
ASSERT_OK(Flush(cf));
TablePropertiesCollection fname_to_props;
ASSERT_OK(db_->GetPropertiesOfAllTables(handles_[cf], &fname_to_props));
ASSERT_EQ(1U, fname_to_props.size());
std::string expected_cf_name;
if (cf > 0) {
expected_cf_name = kExtraCfName;
} else {
expected_cf_name = kDefaultColumnFamilyName;
}
ASSERT_EQ(expected_cf_name,
fname_to_props.begin()->second->column_family_name);
ASSERT_EQ(cf, static_cast<uint32_t>(
fname_to_props.begin()->second->column_family_id));
}
}
TEST_F(DBTablePropertiesTest, GetDbIdentifiersProperty) {
CreateAndReopenWithCF({"goku"}, CurrentOptions());
for (uint32_t cf = 0; cf < 2; ++cf) {
ASSERT_OK(Put(cf, "key", "val"));
ASSERT_OK(Put(cf, "foo", "bar"));
ASSERT_OK(Flush(cf));
TablePropertiesCollection fname_to_props;
ASSERT_OK(db_->GetPropertiesOfAllTables(handles_[cf], &fname_to_props));
ASSERT_EQ(1U, fname_to_props.size());
std::string id, sid;
ASSERT_OK(db_->GetDbIdentity(id));
ASSERT_OK(db_->GetDbSessionId(sid));
ASSERT_EQ(id, fname_to_props.begin()->second->db_id);
ASSERT_EQ(sid, fname_to_props.begin()->second->db_session_id);
}
}
class DBTableHostnamePropertyTest
: public DBTestBase,
public ::testing::WithParamInterface<std::tuple<int, std::string>> {
public:
DBTableHostnamePropertyTest()
: DBTestBase("db_table_hostname_property_test",
/*env_do_fsync=*/false) {}
};
TEST_P(DBTableHostnamePropertyTest, DbHostLocationProperty) {
option_config_ = std::get<0>(GetParam());
Options opts = CurrentOptions();
std::string expected_host_id = std::get<1>(GetParam());
;
if (expected_host_id == kHostnameForDbHostId) {
ASSERT_OK(env_->GetHostNameString(&expected_host_id));
} else {
opts.db_host_id = expected_host_id;
}
CreateAndReopenWithCF({"goku"}, opts);
for (uint32_t cf = 0; cf < 2; ++cf) {
ASSERT_OK(Put(cf, "key", "val"));
ASSERT_OK(Put(cf, "foo", "bar"));
ASSERT_OK(Flush(cf));
TablePropertiesCollection fname_to_props;
ASSERT_OK(db_->GetPropertiesOfAllTables(handles_[cf], &fname_to_props));
ASSERT_EQ(1U, fname_to_props.size());
ASSERT_EQ(fname_to_props.begin()->second->db_host_id, expected_host_id);
}
}
INSTANTIATE_TEST_CASE_P(
DBTableHostnamePropertyTest, DBTableHostnamePropertyTest,
::testing::Values(
// OptionConfig, override db_host_location
std::make_tuple(DBTestBase::OptionConfig::kDefault,
kHostnameForDbHostId),
std::make_tuple(DBTestBase::OptionConfig::kDefault, "foobar"),
std::make_tuple(DBTestBase::OptionConfig::kDefault, ""),
std::make_tuple(DBTestBase::OptionConfig::kPlainTableFirstBytePrefix,
kHostnameForDbHostId),
std::make_tuple(DBTestBase::OptionConfig::kPlainTableFirstBytePrefix,
"foobar"),
std::make_tuple(DBTestBase::OptionConfig::kPlainTableFirstBytePrefix,
"")));
class DeletionTriggeredCompactionTestListener : public EventListener {
public:
void OnCompactionBegin(DB*, const CompactionJobInfo& ci) override {
ASSERT_EQ(ci.compaction_reason,
CompactionReason::kFilesMarkedForCompaction);
}
void OnCompactionCompleted(DB*, const CompactionJobInfo& ci) override {
ASSERT_EQ(ci.compaction_reason,
CompactionReason::kFilesMarkedForCompaction);
}
};
TEST_P(DBTablePropertiesTest, DeletionTriggeredCompactionMarking) {
int kNumKeys = 1000;
int kWindowSize = 100;
int kNumDelsTrigger = 90;
std::shared_ptr<TablePropertiesCollectorFactory> compact_on_del =
NewCompactOnDeletionCollectorFactory(kWindowSize, kNumDelsTrigger);
Options opts = CurrentOptions();
opts.statistics = ROCKSDB_NAMESPACE::CreateDBStatistics();
opts.table_properties_collector_factories.emplace_back(compact_on_del);
if (GetParam() == "kCompactionStyleUniversal") {
opts.compaction_style = kCompactionStyleUniversal;
}
Reopen(opts);
// add an L1 file to prevent tombstones from dropping due to obsolescence
// during flush
ASSERT_OK(Put(Key(0), "val"));
ASSERT_OK(Flush());
MoveFilesToLevel(1);
DeletionTriggeredCompactionTestListener* listener =
new DeletionTriggeredCompactionTestListener();
opts.listeners.emplace_back(listener);
Reopen(opts);
for (int i = 0; i < kNumKeys; ++i) {
if (i >= kNumKeys - kWindowSize &&
i < kNumKeys - kWindowSize + kNumDelsTrigger) {
ASSERT_OK(Delete(Key(i)));
} else {
ASSERT_OK(Put(Key(i), "val"));
}
}
ASSERT_OK(Flush());
ASSERT_OK(dbfull()->TEST_WaitForCompact());
ASSERT_EQ(0, NumTableFilesAtLevel(0));
// Change the window size and deletion trigger and ensure new values take
// effect
kWindowSize = 50;
kNumDelsTrigger = 40;
static_cast<CompactOnDeletionCollectorFactory*>(compact_on_del.get())
->SetWindowSize(kWindowSize);
static_cast<CompactOnDeletionCollectorFactory*>(compact_on_del.get())
->SetDeletionTrigger(kNumDelsTrigger);
for (int i = 0; i < kNumKeys; ++i) {
if (i >= kNumKeys - kWindowSize &&
i < kNumKeys - kWindowSize + kNumDelsTrigger) {
ASSERT_OK(Delete(Key(i)));
} else {
ASSERT_OK(Put(Key(i), "val"));
}
}
ASSERT_OK(Flush());
ASSERT_OK(dbfull()->TEST_WaitForCompact());
ASSERT_EQ(0, NumTableFilesAtLevel(0));
// Change the window size to disable delete triggered compaction
kWindowSize = 0;
static_cast<CompactOnDeletionCollectorFactory*>(compact_on_del.get())
->SetWindowSize(kWindowSize);
static_cast<CompactOnDeletionCollectorFactory*>(compact_on_del.get())
->SetDeletionTrigger(kNumDelsTrigger);
for (int i = 0; i < kNumKeys; ++i) {
if (i >= kNumKeys - kWindowSize &&
i < kNumKeys - kWindowSize + kNumDelsTrigger) {
ASSERT_OK(Delete(Key(i)));
} else {
ASSERT_OK(Put(Key(i), "val"));
}
}
ASSERT_OK(Flush());
ASSERT_OK(dbfull()->TEST_WaitForCompact());
ASSERT_EQ(1, NumTableFilesAtLevel(0));
ASSERT_LT(0, opts.statistics->getTickerCount(COMPACT_WRITE_BYTES_MARKED));
ASSERT_LT(0, opts.statistics->getTickerCount(COMPACT_READ_BYTES_MARKED));
}
TEST_P(DBTablePropertiesTest, RatioBasedDeletionTriggeredCompactionMarking) {
constexpr int kNumKeys = 1000;
constexpr int kWindowSize = 0;
constexpr int kNumDelsTrigger = 0;
constexpr double kDeletionRatio = 0.1;
std::shared_ptr<TablePropertiesCollectorFactory> compact_on_del =
NewCompactOnDeletionCollectorFactory(kWindowSize, kNumDelsTrigger,
kDeletionRatio);
Options opts = CurrentOptions();
opts.statistics = ROCKSDB_NAMESPACE::CreateDBStatistics();
opts.table_properties_collector_factories.emplace_back(compact_on_del);
Reopen(opts);
// Add an L2 file to prevent tombstones from dropping due to obsolescence
// during flush
ASSERT_OK(Put(Key(0), "val"));
ASSERT_OK(Flush());
MoveFilesToLevel(2);
auto* listener = new DeletionTriggeredCompactionTestListener();
opts.listeners.emplace_back(listener);
Reopen(opts);
// Generate one L0 with kNumKeys Put.
for (int i = 0; i < kNumKeys; ++i) {
ASSERT_OK(Put(Key(i), "not important"));
}
ASSERT_OK(Flush());
// Generate another L0 with kNumKeys Delete.
// This file, due to deletion ratio, will trigger compaction: 2@0 files to L1.
// The resulting L1 file has only one tombstone for user key 'Key(0)'.
// Again, due to deletion ratio, a compaction will be triggered: 1@1 + 1@2
// files to L2. However, the resulting file is empty because the tombstone
// and value are both dropped.
for (int i = 0; i < kNumKeys; ++i) {
ASSERT_OK(Delete(Key(i)));
}
ASSERT_OK(Flush());
ASSERT_OK(dbfull()->TEST_WaitForCompact());
for (int i = 0; i < 3; ++i) {
ASSERT_EQ(0, NumTableFilesAtLevel(i));
}
}
INSTANTIATE_TEST_CASE_P(DBTablePropertiesTest, DBTablePropertiesTest,
::testing::Values("kCompactionStyleLevel",
"kCompactionStyleUniversal"));
} // namespace ROCKSDB_NAMESPACE
int main(int argc, char** argv) {
ROCKSDB_NAMESPACE::port::InstallStackTraceHandler();
::testing::InitGoogleTest(&argc, argv);
return RUN_ALL_TESTS();
}