Summary:
The old cleanup code had a race condition:
1. Test thread: DestroyDB() marked a file as trash
2. DeleteScheduler thread: Got the file's size and decided to delete it in chunks
3. Test thread: DestroyDir() deleted that trash file
4. DeleteScheduler thread: Began deleting in chunks starting by calling ReopenWritableFile(). Unfortunately this recreates the deleted trash file
5. Test thread: DestroyDir() fails to remove the parent directory because it contains the file created in 4.
6. Test thread: Checkpoint::Create() fails due to the directory already existing
It could be repro'd with the following patch/command.
Patch:
```
diff --git a/file/delete_scheduler.cc b/file/delete_scheduler.cc
index 8a2d1615d..337d24a60 100644
--- a/file/delete_scheduler.cc
+++ b/file/delete_scheduler.cc
@@ -317,6 +317,12 @@ Status DeleteScheduler::DeleteTrashFile(const std::string& path_in_trash,
&num_hard_links, nullptr);
if (my_status.ok()) {
if (num_hard_links == 1) {
+ // Give some time for DestroyDir() to delete file entries. Then, the
+ // below `ReopenWritableFile()` will recreate files, preventing the
+ // parent directory from being deleted.
+ if (rand() % 2 == 0) {
+ usleep(1000);
+ }
std::unique_ptr<FSWritableFile> wf;
my_status = fs_->ReopenWritableFile(path_in_trash, FileOptions(), &wf,
nullptr);
diff --git a/file/file_util.cc b/file/file_util.cc
index 43608fcdc..2cee1ad8e 100644
--- a/file/file_util.cc
+++ b/file/file_util.cc
@@ -263,6 +263,13 @@ Status DestroyDir(Env* env, const std::string& dir) {
}
}
+ // Give some time for the DeleteScheduler thread's ReopenWritableFile() to
+ // recreate deleted files
+ if (dir.find("checkpoint") != std::string::npos) {
+ fprintf(stderr, "waiting to destroy %s\n", dir.c_str());
+ usleep(10000);
+ }
+
if (s.ok()) {
s = env->DeleteDir(dir);
// DeleteDir might or might not report NotFound
```
Command:
```
TEST_TMPDIR=/dev/shm python3 tools/db_crashtest.py blackbox --simple --write_buffer_size=131072 --target_file_size_base=131072 --max_bytes_for_level_base=524288 --checkpoint_one_in=100 --clear_column_family_one_in=0 --max_key=1000 --value_size_mult=33 --sst_file_manager_bytes_per_truncate=4096 --sst_file_manager_bytes_per_sec=1048576 --interval=3 --compression_type=none --sync_fault_injection=1
```
Obviously we don't want to use scheduled deletion here as we need the checkpoint directory deleted immediately. I suspect the DestroyDir() was an attempt to fixup incomplete DestroyDB()s. Now that we expect DestroyDB() to be complete I removed that code.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/11389
Reviewed By: hx235
Differential Revision: D45137142
Pulled By: ajkr
fbshipit-source-id: 2af743d342c77cc414fd25fc4c9d7c9c6079ad24
Summary:
during manual compaction (CompactRange()), L0->L1 trivial move is disabled when only L0 overlaps with compacting key range (introduced in https://github.com/facebook/rocksdb/issues/7368 to enforce kForce* contract). This can cause large memory usage due to compaction readahead when number of L0 files is large. This PR allows L0->L1 trivial move in this case, and will do a L1 -> L1 intra-level compaction when needed (`bottommost_level_compaction` is kForce*). In brief, consider a DB with only L0 file, and user calls CompactRange(kForce, nullptr, nullptr),
- before this PR, RocksDB does a L0 -> L1 compaction (disallow trivial move),
- after this PR, RocksDB does a L0 -> L1 compaction (allow trivial move), and a L1 -> L1 compaction.
Users can use kForceOptimized to avoid this extra L1->L1 compaction overhead when L0s are overlapping and cannot be trivial moved.
This PR also fixed a bug (see previous discussion in https://github.com/facebook/rocksdb/issues/11041) where `final_output_level` of a manual compaction can be miscalculated when `level_compaction_dynamic_level_bytes=true`. This bug could cause incorrect level being moved when CompactRangeOptions::change_level is specified.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/11375
Test Plan: - Added new unit tests to test that L0 -> L1 compaction allows trivial move and L1 -> L1 compaction is done when needed.
Reviewed By: ajkr
Differential Revision: D44943518
Pulled By: cbi42
fbshipit-source-id: e9fb770d17b163c18a623e1d1bd6b81159192708
Summary:
Because of this failure with snappy 1.1.8, ROCKSDB_NO_FBCODE=1
```
Value 3531 is not in range [2000, 3525]
table/table_test.cc:4231: Failure
```
Pull Request resolved: https://github.com/facebook/rocksdb/pull/11384
Test Plan: run updated test in failing configuration
Reviewed By: ajkr
Differential Revision: D45057161
Pulled By: pdillinger
fbshipit-source-id: 397054f08033315e2e2bd9410f1fa32ddbf3b9c8
Summary:
This test exhibited the following flaky failure:
```
db/db_write_test.cc:653: Failure
db_->Resume()
Corruption: Not active
```
I was able to repro it by applying the following patch to coerce a specific race condition:
```
diff --git a/db/db_write_test.cc b/db/db_write_test.cc
index d82c57376..775ba3cde 100644
--- a/db/db_write_test.cc
+++ b/db/db_write_test.cc
@@ -636,6 +636,10 @@ TEST_P(DBWriteTest, LockWALInEffect) {
ASSERT_TRUE(dbfull()->WALBufferIsEmpty());
ASSERT_OK(db_->UnlockWAL());
+ // Test thread: sleep interval: [0, 3)
+ // In this interval, the file system is active
+ sleep(3);
+
// Fail the WAL flush if applicable
fault_fs->SetFilesystemActive(false);
Status s = Put("key2", "value");
@@ -649,6 +653,11 @@ TEST_P(DBWriteTest, LockWALInEffect) {
ASSERT_OK(db_->LockWAL());
ASSERT_OK(db_->UnlockWAL());
}
+
+ // Test thread: sleep interval: [3, 6)
+ // In this interval, the file system is inactive
+ sleep(3);
+
fault_fs->SetFilesystemActive(true);
ASSERT_OK(db_->Resume());
// Writes should work again
diff --git a/db/flush_job.cc b/db/flush_job.cc
index 8193f594f..602ee2c9f 100644
--- a/db/flush_job.cc
+++ b/db/flush_job.cc
@@ -979,6 +979,10 @@ Status FlushJob::WriteLevel0Table() {
DirFsyncOptions(DirFsyncOptions::FsyncReason::kNewFileSynced));
}
TEST_SYNC_POINT_CALLBACK("FlushJob::WriteLevel0Table", &mems_);
+ // Flush thread: sleep interval: [0, 4)
+ // Upon awakening, the file system will be inactive. Then the MANIFEST
+ // update will fail.
+ sleep(4);
db_mutex_->Lock();
}
base_->Unref();
```
The fix for this scenario is explained in the code change.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/11382
Reviewed By: cbi42
Differential Revision: D45027632
Pulled By: ajkr
fbshipit-source-id: 6bfa35a5781c0c080fb74e13f2b2c9f871f7effb
Summary:
In CircleCI build-linux-arm-test-full job (https://app.circleci.com/pipelines/github/facebook/rocksdb/26462/workflows/a9d39d2c-c970-4b0f-9c10-7743beb9771b/jobs/591722), this test exhibited the following flaky failure:
```
db/db_bloom_filter_test.cc:2506: Failure
Expected: (TestGetTickerCount(options, BLOOM_FILTER_USEFUL)) > (65000 * 2), actual: 120558 vs 130000
```
I ssh'd to an instance and observed it cuts memtables at slightly different points across runs. Logging in `ConcurrentArena` pointed to `try_lock()` returning false at different points across runs.
This PR changes the approach to allow a fixed number of keys per memtable flush. I verified the bloom filter useful count is deterministic now even on the CircleCI ARM instance.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/11383
Reviewed By: cbi42
Differential Revision: D45036829
Pulled By: ajkr
fbshipit-source-id: b602dacb63955f1af09bf0ed409cde0552805a08
Summary:
When calculating the largest_key in ImportColumnFamilyJob::GetIngestedFileInfo, only the first element of range_del_iter is calculated. If range_del_iter has multiple elements, the largest_key will be wrong
Pull Request resolved: https://github.com/facebook/rocksdb/pull/11381
Reviewed By: cbi42
Differential Revision: D44981450
Pulled By: ajkr
fbshipit-source-id: 584bc7da86295568a96984d2951644f289e578c7
Summary:
Before this PR, in `LevelCompactionBuilder::TryExtendNonL0TrivialMove(index)`, we start from a file at index and expand the compaction input towards right to find files to trivial move. This PR adds the logic to also expand towards left.
Another major change made in this PR is to not expand L0 files through `TryExtendNonL0TrivialMove()`. This happens currently when compacting L0 files to an empty output level. The condition for expanding files in `TryExtendNonL0TrivialMove()` is to check atomic boundary, which does not take into account that L0 files can overlap in key range and are not sorted in key order. So it may include more L0 files than needed and disallow a trivial move. This change is included in this PR so that we don't make it worse by always expanding L0 in both direction.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/11347
Test Plan:
* new unit test
* Benchmark does not show obvious improvement or regression:
```
Write sequentially
./db_bench --benchmarks=fillseq --compression_type=lz4 --write_buffer_size=1000000 --num=100000000 --value_size=100 -level_compaction_dynamic_level_bytes --target_file_size_base=7340032 --max_bytes_for_level_base=16777216
Main:
fillseq : 4.726 micros/op 211592 ops/sec 472.607 seconds 100000000 operations; 23.4 MB/s
This PR:
fillseq : 4.755 micros/op 210289 ops/sec 475.534 seconds 100000000 operations; 23.3 MB/s
Write randomly
./db_bench --benchmarks=fillrandom --compression_type=lz4 --write_buffer_size=1000000 --num=100000000 --value_size=100 -level_compaction_dynamic_level_bytes --target_file_size_base=7340032 --max_bytes_for_level_base=16777216
Main:
fillrandom : 16.351 micros/op 61159 ops/sec 1635.066 seconds 100000000 operations; 6.8 MB/s
This PR:
fillrandom : 15.798 micros/op 63298 ops/sec 1579.817 seconds 100000000 operations; 7.0 MB/s
```
Reviewed By: ajkr
Differential Revision: D44645650
Pulled By: cbi42
fbshipit-source-id: 8631f3a6b3f01decbbf18c34f2b62833cb4f9733
Summary:
**Context/Summary:**
ASSERT_EQ will only verify the code of Status, but will not check the state message of Status.
- Assert by checking Status state in `ImportColumnFamilyTest`
- Forgot to set db_comparator_name when creating ExportImportFilesMetaData in `ImportColumnFamilyNegativeTest`
Pull Request resolved: https://github.com/facebook/rocksdb/pull/11372
Reviewed By: ajkr
Differential Revision: D45004343
Pulled By: cbi42
fbshipit-source-id: a13d45521df17ead3d6d4c1c1fe1e4c95397ce8b
Summary:
util/ribbon_test.cc: avoid ambiguous reversed operator error in c++20 (and enable checking for the error)
Code would produce errors like this, when compiled with -Wambiguous-reversed-operator under c++20.
```
util/ribbon_test.cc:695:20: error: ISO C++20 considers use of overloaded operator '!=' (with operand types 'KeyGen' (aka '(anonymous namespace)::StandardKeyGen') and 'KeyGen') to be ambiguou
s despite there being a unique best viable function with non-reversed arguments [-Werror,-Wambiguous-reversed-operator]
while (cur != batch_end) {
~~~ ^ ~~~~~~~~~
util/ribbon_test.cc:111:8: note: candidate function with non-reversed arguments
bool operator!=(const StandardKeyGen& other) {
^
util/ribbon_test.cc:107:8: note: ambiguous candidate function with reversed arguments
bool operator==(const StandardKeyGen& other) {
^
```
This will become a hard error in future standards.
Confirmed that no errors were generated when building using clang and c++20:
```
USE_CLANG=1 USE_COROUTINES=1 make
```
Pull Request resolved: https://github.com/facebook/rocksdb/pull/11371
Reviewed By: meyering
Differential Revision: D44921027
Pulled By: cbi42
fbshipit-source-id: ef25b78260920a4d75a718310688d3a2487ffa87
Summary:
This option is immutable through the life time of the DB open. For now, updating its value between different DB open sessions is also a non compatible change. When I work on support for updating comparator, the type of updates accepted for this option will be supported then.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/11362
Test Plan: `make check`
Reviewed By: ltamasi
Differential Revision: D44873870
Pulled By: jowlyzhang
fbshipit-source-id: aa02094754b58d99abf9af4c9a8108c1350254cb
Summary:
Makes it easier to use generated Rust bindings. Constness of these is already part of the C++ API.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/11243
Reviewed By: hx235
Differential Revision: D44840394
Pulled By: ajkr
fbshipit-source-id: bcd1aeb8c959c304148d25b00043bb8c4cd3e0a4
Summary:
The CI systems other than CircleCI are almost always in a failing state. Since CircleCI covers linux, macos, and windows, we can remove the others.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/11354
Reviewed By: hx235
Differential Revision: D44774627
Pulled By: ajkr
fbshipit-source-id: c83b298ec5afe4ea410744eda6cc98fc6a3365f1
Summary:
After https://github.com/facebook/rocksdb/issues/11301, I wasn't sure whether I had regressed block cache tracing with MultiGet. Demo PR https://github.com/facebook/rocksdb/issues/11330 shows the flawed state of tracing MultiGet before my change, and based on the unit test, there was essentially no change in tracing behavior with https://github.com/facebook/rocksdb/issues/11301. This change is to leave that code and behavior better than I found it.
This change is not intended to change any production behaviors except when block cache tracing is active, though might improve general read path efficiency by disabling some related tracking when such tracing is disabled.
More detail on production code:
* Refactoring to consolidate the construction of BlockCacheTraceRecord, and other related functionality, in block-based table reader, though it's somewhat awkward to preserve an optimization to avoid copying Slices into temporary strings in BlockCacheLookupContext.
* Accurately track cache hits and misses (etc.) for each data block accessed by a MultiGet(). (Previously reported hits as misses.)
* Reduced repeated checking of `block_cache_tracer_` state (by creating lookup_context only when active) for efficiency and to reduce the risk of corner case bugs where tracing is enabled or disabled for different parts of a read op. (See a TODO below)
* Improved estimate calculation for num_keys_in_block (see code comment)
Possible follow-up:
* `XXX:` use_cache=true means double cache query? (possible double-query of block cache when allow_mmap_reads=true)
* `TODO:` need more than one lookup_context here to track individual filter and index partition hits and misses
* `TODO:` optimize more state checks of `block_cache_tracer_` down to `lookup_context != nullptr`
* Pre-existing `XXX:` There appear to be 'break' statements above that bypass this writing of the block cache trace record
* Expand test coverage (see below)
Pull Request resolved: https://github.com/facebook/rocksdb/pull/11339
Test Plan:
* Added a basic unit test for block cache tracing MultiGet, for now just covering one data block with two keys.
* Added HitMissCountingCache to independently verify that the actual block cache trace and expected block cache trace also agree with the actual number of cache hits / misses (nothing missing or mislabeled). For now only used with MultiGet test.
* Better testing of num_keys_in_block, for now just with MultiGet
* Misc improvements to table_test to improve clarity, such as making it clear that certain keys are auto-inserted at the start of every test.
Performance test:
Testing multireadrandom as in https://github.com/facebook/rocksdb/issues/11301, except averaging over distinct runs rather than [-X30] which doesn't seem to sufficiently reset after each run to work as an independent test run.
Base with revert of 11301: 3148926 ops/sec
Base: 3019146 ops/sec
New: 2999529 ops/sec
Possibly a tiny MultiGet CPU regression with this change. We are now always allocating an additional vector for the LookupContexts. I'm still contemplating options to try to correct the regression in https://github.com/facebook/rocksdb/issues/11301.
Testing readrandom:
Base with revert of 11301: 2311988
Base: 2281726
New: 2299722
Possibly a tiny Get CPU improvement with this change. We are now avoiding some unnecessary LookupContext population.
Reviewed By: akankshamahajan15
Differential Revision: D44557845
Pulled By: pdillinger
fbshipit-source-id: b841691799d2a48fb59cc8880dc7cbb1e107ae3d
Summary:
If RocksDB enables user-defined timestamp, then RocksDB read path can filter table files by the min/max timestamps of each file. If application wants to lookup a key that is the most recent and visible to a certain timestamp ts, then we can compare ts with the min_ts of each file. If ts < min_ts, then we know all keys in the file is not visible at time ts, then we do not have to open the file. This can also save an IO.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/11332
Reviewed By: pdillinger
Differential Revision: D44763497
Pulled By: guowentian
fbshipit-source-id: abde346b9f18480fe03c04e4006e7d62aa9c22a8
Summary:
When a user migrates to level compaction + `level_compaction_dynamic_level_bytes=true`, or when a DB shrinks, there can be unnecessary levels in the DB. Before this PR, this is no way to remove these levels except a manual compaction. These extra unnecessary levels make it harder to guarantee max_bytes_for_level_multiplier and can cause extra space amp. This PR boosts compaction score for these levels to allow RocksDB to automatically drain these levels. Together with https://github.com/facebook/rocksdb/issues/11321, this makes migration to `level_compaction_dynamic_level_bytes=true` automatic without needing user to do a one time full manual compaction. Credit: this PR is modified from https://github.com/facebook/rocksdb/issues/3921.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/11340
Test Plan:
- New unit tests
- `python3 tools/db_crashtest.py whitebox --simple` which randomly sets level_compaction_dynamic_level_bytes in each run.
Reviewed By: ajkr
Differential Revision: D44563884
Pulled By: cbi42
fbshipit-source-id: e20d3620bd73dff22be18c5a91a07f340740bcc8
Summary:
VerifyFileChecksums currently interprets the readahead_size as a payload of readahead_size for calculating the checksum, plus a prefetch of an additional readahead_size. Hence each read is readahead_size * 2. This change treats it as chunks of readahead_size for checksum calculation.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/11328
Test Plan: Add a unit test
Reviewed By: pdillinger
Differential Revision: D44718781
Pulled By: anand1976
fbshipit-source-id: 79bae1ebaa27de2a13bc86f5910bf09356936e63
Summary:
I previously misread or misinterpreted API contracts for SecondaryCache and this should correct the record. (Follow-up item from https://github.com/facebook/rocksdb/issues/11301)
Pull Request resolved: https://github.com/facebook/rocksdb/pull/11316
Test Plan: comments only
Reviewed By: anand1976
Differential Revision: D44245107
Pulled By: pdillinger
fbshipit-source-id: 3f8ddec150674b75728f1730f99b963bbf7b76e7
Summary:
... which increases default number of shards from 16 to 64. Although the default block cache size is only recommended for applications where RocksDB is not performance-critical, under stress conditions, block cache mutex contention could become a performance bottleneck. This change of default should alleviate that.
Note that reducing the size of cache shards (recommended minimum 512MB) could cause thrashing, e.g. on filter blocks, so capacity needs to increase to safely increase number of shards.
The 8MB default dates back to 2011 or earlier (f779e7a5), when the most simultaneous threads you could get from a single CPU socket was 20 (e.g. Intel Xeon E7-8870). Now more than 100 is available.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/11350
Test Plan: unit tests updated
Reviewed By: cbi42
Differential Revision: D44674873
Pulled By: pdillinger
fbshipit-source-id: 91ed3070789b42679283c7e6dc97c41a6a97bdf4
Summary:
**Context/Summary:**
- Allow runtime changes to whether `WriteBufferManager` allows stall or not by calling `SetAllowStall()`
- Misc: some clean up - see PR conversation
Pull Request resolved: https://github.com/facebook/rocksdb/pull/11335
Test Plan: - New UT
Reviewed By: akankshamahajan15
Differential Revision: D44502555
Pulled By: hx235
fbshipit-source-id: 24b5cc57df7734b11d42e4870c06c87b95312b5e
Summary:
Similarly to `GetEntity` prior to https://github.com/facebook/rocksdb/issues/11303, the `MultiGetEntity` API is currently
only used in the DB verification logic of the stress tests. The patch introduces
a new mode where all point lookups are performed using `MultiGetEntity`,
and implements the corresponding logic in the non-batched, batched, and
CF consistency tests.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/11336
Test Plan: Ran simple blackbox tests for the various stress test flavors.
Reviewed By: akankshamahajan15
Differential Revision: D44513285
Pulled By: ltamasi
fbshipit-source-id: c3db098501bf875b6a356b09fc676a0268d92c35
Summary:
**Context/Summary:**
Motived by user need of investigating db iterator behavior during an interval of any time length of a certain thread, we decide to collect and expose related counters in `PerfContext` as an experimental feature, in addition to the existing db-scope ones (i.e, tickers)
Pull Request resolved: https://github.com/facebook/rocksdb/pull/11320
Test Plan:
- new UT
- db bench
Setup
```
./db_bench -db=/dev/shm/testdb/ -benchmarks="fillseq" -key_size=32 -value_size=512 -num=1000000 -compression_type=none -bloom_bits=3
```
Test till converges
```
./db_bench -seed=1679526311157283 -use_existing_db=1 -perf_level=2 -db=/dev/shm/testdb/ -benchmarks="seekrandom[-X60]"
```
pre-change
`seekrandom [AVG 33 runs] : 7545 (± 100) ops/sec`
post-change (no regression)
`seekrandom [AVG 33 runs] : 7688 (± 67) ops/sec`
Reviewed By: cbi42
Differential Revision: D44321931
Pulled By: hx235
fbshipit-source-id: f98a254ba3e3ced95eb5928884e33f1b99dca401
Summary:
…evel_bytes
During DB open, if a column family uses level compaction with level_compaction_dynamic_level_bytes=true, trivially move its files down in the LSM such that the bottommost files are in Lmax, the second from bottommost level files are in Lmax-1 and so on. This is aimed to make it easier to migrate level_compaction_dynamic_level_bytes from false to true. Before this change, a full manual compaction is suggested for such migration. After this change, user can just restart DB to turn on this option. db_crashtest.py is updated to randomly choose value for level_compaction_dynamic_level_bytes.
Note that there may still be too many unnecessary levels if a user is migrating from universal compaction or level compaction with a smaller level multiplier. A full manual compaction may still be needed in that case before some PR that automatically drain unnecessary levels like https://github.com/facebook/rocksdb/issues/3921 lands. Eventually we may want to change the default value of option level_compaction_dynamic_level_bytes to true.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/11321
Test Plan:
1. Added unit tests.
2. Crash test: ran a variation of db_crashtest.py (like 32516507e77521ae887e45091b69139e32e8efb7) that turns level_compaction_dynamic_level_bytes on and off and switches between LC and UC for the same DB.
TODO: Update `OptionChangeMigration`, either after this PR or https://github.com/facebook/rocksdb/issues/3921.
Reviewed By: ajkr
Differential Revision: D44341930
Pulled By: cbi42
fbshipit-source-id: 013de19a915c6a0502be569f07c4cc8f1c3c6be2
Summary:
This include is unused in the header. In one build environment of ours, stdarg.h is actually not present, and this include prevents us from building rocksdb dependencies.
We're currently monkey-patching this line out in our build script (still WIP), which of course is not good. ec2852caa3
Note that removing this include might break builds in unexpected ways that include rocksdb/c.h and then use `va_start`, `va_end`, etc. However, if you're using these functions, you really should include stdarg.h yourself, so I don't think this should prevent this PR.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/11302
Reviewed By: ajkr
Differential Revision: D44139819
Pulled By: cbi42
fbshipit-source-id: 10c40b0b0260b23ccb7dc84e55a993c7dfbdc4cf
Summary:
In DBCompactionTest::CancelCompactionWaitingOnConflict, when generating SST files to trigger a compaction, we don't wait after each file, which may cause multiple memtables going to the same SST file, causing insufficient files to trigger the compaction. We do the waiting instead, except the last one, which would trigger compaction.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/11318
Test Plan: Run DBCompactionTest.CancelCompactionWaitingOnConflict multiple times.
Reviewed By: ajkr
Differential Revision: D44267273
fbshipit-source-id: 86af49b05fc67ea3335312f0f5f3d22df1520bf8
Summary:
Right now, EnvLogger has the same IO error assertion as most other places: if we are writing to the file after we've seen an IO error, the assertion would trigger. This is too strict for info logger: we would not fail DB if info logger fails and we would try the best to continue logging. For now, we simplify the problem by disabling the assertion for EnvLogger.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/11314
Test Plan: Run env_logger_test to make sure at least it doesn't fail in normal cases.
Reviewed By: anand1976
Differential Revision: D44227732
fbshipit-source-id: e3d31a221a5757f018a67ccaa96dcf89eb981f66
Summary:
Internally refactors SecondaryCache integration out of LRUCache specifically and into a wrapper/adapter class that works with various Cache implementations. Notably, this relies on separating the notion of async lookup handles from other cache handles, so that HyperClockCache doesn't have to deal with the problem of allocating handles from the hash table for lookups that might fail anyway, and might be on the same key without support for coalescing. (LRUCache's hash table can incorporate previously allocated handles thanks to its pointer indirection.) Specifically, I'm worried about the case in which hundreds of threads try to access the same block and probing in the hash table degrades to linear search on the pile of entries with the same key.
This change is a big step in the direction of supporting stacked SecondaryCaches, but there are obstacles to completing that. Especially, there is no SecondaryCache hook for evictions to pass from one to the next. It has been proposed that evictions be transmitted simply as the persisted data (as in SaveToCallback), but given the current structure provided by the CacheItemHelpers, that would require an extra copy of the block data, because there's intentionally no way to ask for a contiguous Slice of the data (to allow for flexibility in storage). `AsyncLookupHandle` and the re-worked `WaitAll()` should be essentially prepared for stacked SecondaryCaches, but several "TODO with stacked secondaries" issues remain in various places.
It could be argued that the stacking instead be done as a SecondaryCache adapter that wraps two (or more) SecondaryCaches, but at least with the current API that would require an extra heap allocation on SecondaryCache Lookup for a wrapper SecondaryCacheResultHandle that can transfer a Lookup between secondaries. We could also consider trying to unify the Cache and SecondaryCache APIs, though that might be difficult if `AsyncLookupHandle` is kept a fixed struct.
## cache.h (public API)
Moves `secondary_cache` option from LRUCacheOptions to ShardedCacheOptions so that it is applicable to HyperClockCache.
## advanced_cache.h (advanced public API)
* Add `Cache::CreateStandalone()` so that the SecondaryCache support wrapper can use it.
* Add `SetEvictionCallback()` / `eviction_callback_` so that the SecondaryCache support wrapper can use it. Only a single callback is supported for efficiency. If there is ever a need for more than one, hopefully that can be handled with a broadcast callback wrapper.
These are essentially the two "extra" pieces of `Cache` for pulling out specific SecondaryCache support from the `Cache` implementation. I think it's a good trade-off as these are reasonable, limited, and reusable "cut points" into the `Cache` implementations.
* Remove async capability from standard `Lookup()` (getting rid of awkward restrictions on pending Handles) and add `AsyncLookupHandle` and `StartAsyncLookup()`. As noted in the comments, the full struct of `AsyncLookupHandle` is exposed so that it can be stack allocated, for efficiency, though more data is being copied around than before, which could impact performance. (Lookup info -> AsyncLookupHandle -> Handle vs. Lookup info -> Handle)
I could foresee a future in which a Cache internally saves a pointer to the AsyncLookupHandle, which means it's dangerous to allow it to be copyable or even movable. It also means it's not compatible with std::vector (which I don't like requiring as an API parameter anyway), so `WaitAll()` expects any contiguous array of AsyncLookupHandles. I believe this is best for common case efficiency, while behaving well in other cases also. For example, `WaitAll()` has no effect on default-constructed AsyncLookupHandles, which look like a completed cache miss.
## cacheable_entry.h
A couple of functions are obsolete because Cache::Handle can no longer be pending.
## cache.cc
Provides default implementations for new or revamped Cache functions, especially appropriate for non-blocking caches.
## secondary_cache_adapter.{h,cc}
The full details of the Cache wrapper adding SecondaryCache support. Essentially replicates the SecondaryCache handling that was in LRUCache, but obviously refactored. There is a bit of logic duplication, where Lookup() is essentially a manually optimized version of StartAsyncLookup() and Wait(), but it's roughly a dozen lines of code.
## sharded_cache.h, typed_cache.h, charged_cache.{h,cc}, sim_cache.cc
Simply updated for Cache API changes.
## lru_cache.{h,cc}
Carefully remove SecondaryCache logic, implement `CreateStandalone` and eviction handler functionality.
## clock_cache.{h,cc}
Expose existing `CreateStandalone` functionality, add eviction handler functionality. Light refactoring.
## block_based_table_reader*
Mostly re-worked the only usage of async Lookup, which is in BlockBasedTable::MultiGet. Used arrays in place of autovector in some places for efficiency. Simplified some logic by not trying to process some cache results before they're all ready.
Created new function `BlockBasedTable::GetCachePriority()` to reduce some pre-existing code duplication (and avoid making it worse).
Fixed at least one small bug from the prior confusing mixture of async and sync Lookups. In MaybeReadBlockAndLoadToCache(), called by RetrieveBlock(), called by MultiGet() with wait=false, is_cache_hit for the block_cache_tracer entry would not be set to true if the handle was pending after Lookup and before Wait.
## Intended follow-up work
* Figure out if there are any missing stats or block_cache_tracer work in refactored BlockBasedTable::MultiGet
* Stacked secondary caches (see above discussion)
* See if we can make up for the small MultiGet performance regression.
* Study more performance with SecondaryCache
* Items evicted from over-full LRUCache in Release were not being demoted to SecondaryCache, and still aren't to minimize unit test churn. Ideally they would be demoted, but it's an exceptional case so not a big deal.
* Use CreateStandalone for cache reservations (save unnecessary hash table operations). Not a big deal, but worthy cleanup.
* Somehow I got the contract for SecondaryCache::Insert wrong in #10945. (Doesn't take ownership!) That API comment needs to be fixed, but didn't want to mingle that in here.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/11301
Test Plan:
## Unit tests
Generally updated to include HCC in SecondaryCache tests, though HyperClockCache has some different, less strict behaviors that leads to some tests not really being set up to work with it. Some of the tests remain disabled with it, but I think we have good coverage without them.
## Crash/stress test
Updated to use the new combination.
## Performance
First, let's check for regression on caches without secondary cache configured. Adding support for the eviction callback is likely to have a tiny effect, but it shouldn't be worrisome. LRUCache could benefit slightly from less logic around SecondaryCache handling. We can test with cache_bench default settings, built with DEBUG_LEVEL=0 and PORTABLE=0.
```
(while :; do base/cache_bench --cache_type=hyper_clock_cache | grep Rough; done) | awk '{ sum += $9; count++; print $0; print "Average: " int(sum / count) }'
```
**Before** this and #11299 (which could also have a small effect), running for about an hour, before & after running concurrently for each cache type:
HyperClockCache: 3168662 (average parallel ops/sec)
LRUCache: 2940127
**After** this and #11299, running for about an hour:
HyperClockCache: 3164862 (average parallel ops/sec) (0.12% slower)
LRUCache: 2940928 (0.03% faster)
This is an acceptable difference IMHO.
Next, let's consider essentially the worst case of new CPU overhead affecting overall performance. MultiGet uses the async lookup interface regardless of whether SecondaryCache or folly are used. We can configure a benchmark where all block cache queries are for data blocks, and all are hits.
Create DB and test (before and after tests running simultaneously):
```
TEST_TMPDIR=/dev/shm ./db_bench -benchmarks=fillrandom -num=30000000 -disable_wal=1 -bloom_bits=16
TEST_TMPDIR=/dev/shm base/db_bench -benchmarks=multireadrandom[-X30] -readonly -multiread_batched -batch_size=32 -num=30000000 -bloom_bits=16 -cache_size=6789000000 -duration 20 -threads=16
```
**Before**:
multireadrandom [AVG 30 runs] : 3444202 (± 57049) ops/sec; 240.9 (± 4.0) MB/sec
multireadrandom [MEDIAN 30 runs] : 3514443 ops/sec; 245.8 MB/sec
**After**:
multireadrandom [AVG 30 runs] : 3291022 (± 58851) ops/sec; 230.2 (± 4.1) MB/sec
multireadrandom [MEDIAN 30 runs] : 3366179 ops/sec; 235.4 MB/sec
So that's roughly a 3% regression, on kind of a *worst case* test of MultiGet CPU. Similar story with HyperClockCache:
**Before**:
multireadrandom [AVG 30 runs] : 3933777 (± 41840) ops/sec; 275.1 (± 2.9) MB/sec
multireadrandom [MEDIAN 30 runs] : 3970667 ops/sec; 277.7 MB/sec
**After**:
multireadrandom [AVG 30 runs] : 3755338 (± 30391) ops/sec; 262.6 (± 2.1) MB/sec
multireadrandom [MEDIAN 30 runs] : 3785696 ops/sec; 264.8 MB/sec
Roughly a 4-5% regression. Not ideal, but not the whole story, fortunately.
Let's also look at Get() in db_bench:
```
TEST_TMPDIR=/dev/shm ./db_bench -benchmarks=readrandom[-X30] -readonly -num=30000000 -bloom_bits=16 -cache_size=6789000000 -duration 20 -threads=16
```
**Before**:
readrandom [AVG 30 runs] : 2198685 (± 13412) ops/sec; 153.8 (± 0.9) MB/sec
readrandom [MEDIAN 30 runs] : 2209498 ops/sec; 154.5 MB/sec
**After**:
readrandom [AVG 30 runs] : 2292814 (± 43508) ops/sec; 160.3 (± 3.0) MB/sec
readrandom [MEDIAN 30 runs] : 2365181 ops/sec; 165.4 MB/sec
That's showing roughly a 4% improvement, perhaps because of the secondary cache code that is no longer part of LRUCache. But weirdly, HyperClockCache is also showing 2-3% improvement:
**Before**:
readrandom [AVG 30 runs] : 2272333 (± 9992) ops/sec; 158.9 (± 0.7) MB/sec
readrandom [MEDIAN 30 runs] : 2273239 ops/sec; 159.0 MB/sec
**After**:
readrandom [AVG 30 runs] : 2332407 (± 11252) ops/sec; 163.1 (± 0.8) MB/sec
readrandom [MEDIAN 30 runs] : 2335329 ops/sec; 163.3 MB/sec
Reviewed By: ltamasi
Differential Revision: D44177044
Pulled By: pdillinger
fbshipit-source-id: e808e48ff3fe2f792a79841ba617be98e48689f5
Summary:
In PosixFileSystem, IO uring support is opt-in. If the support is not enabled by the user, then ignore the async_io ReadOption in MultiGet and iteration at the top, rather than follow the async_io codepath and transparently switch to sync IO at the FileSystem layer.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/11296
Test Plan: Add new unit tests
Reviewed By: akankshamahajan15
Differential Revision: D44045776
Pulled By: anand1976
fbshipit-source-id: a0881bf763ca2fde50b84063d0068bb521edd8b9
Summary:
The `GetEntity` API is currently used in the stress tests for verification purposes;
this patch extends the coverage by adding a mode where all point lookups in
the non-batched, batched, and CF consistency stress tests are done using this API.
The PR also includes a bit of refactoring to eliminate some boilerplate code around
the wide-column consistency checks.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/11303
Test Plan: Ran stress tests of the batched, non-batched, and CF consistency varieties.
Reviewed By: akankshamahajan15
Differential Revision: D44148503
Pulled By: ltamasi
fbshipit-source-id: fecdbfd3e65a459bbf16ab7aa7b9173e19240077
Summary:
In preparation for factoring secondary cache support out of individual Cache implementations, we can get rid of the "in secondary cache" flag on entries through a workable hack: when an entry is promoted from secondary, it is inserted in primary using a helper that lacks secondary cache support, thus preventing re-insertion into secondary cache through existing logic.
This adds to the complexity of building CacheItemHelpers, because you always have to be able to get to an equivalent helper without secondary cache support, but that complexity is reasonably isolated within RocksDB typed_cache.h and test code.
gcc-7 seems to have problems with constexpr constructor referencing `this` so removed constexpr support on CacheItemHelper.
Also refactored some related test code to share common code / functionality.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/11299
Test Plan: existing tests
Reviewed By: anand1976
Differential Revision: D44101453
Pulled By: pdillinger
fbshipit-source-id: 7a59d0a3938ee40159c90c3e65d7004f6a272345
Andrew Kryczka