Summary:
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9370
GCC and newer clang, e.g. clang-12 treat `std::unique_ptr` slightly differently.
For the following code
```
#include <iostream>
#include <memory>
#include <type_traits>
struct A {
std::unique_ptr<int> m1;
};
int main()
{
std::cout << std::boolalpha;
std::cout << std::is_standard_layout<A>::value << '\n';
return 0;
}
```
GCC11(C++20) (tested on https://en.cppreference.com/w/cpp/types/is_standard_layout) will print "true", while newer clang, e.g. clang-12 will print "false". This breaks the usage of `offsetof()` on structs with non-static members of type `std::unique_ptr`.
Fixing this by replacing the builtin `offsetof` with a trick documented at https://gist.github.com/graphitemaster/494f21190bb2c63c5516.
Reviewed By: jay-zhuang
Differential Revision: D33420840
fbshipit-source-id: 02bde281dfa28809bec787ad0f7019e85dd9c607
Summary:
This change adds the filename of the offending filen to several place that produce Status objects with code `kCorruption`.
This is not an attempt to have every Corruption message in the codebase extended with the filename, but it is a start.
The motivation for the change was to quickly diagnose which file is corrupted when a large database is openend and there is not option to copy it offsite for analysis, run strace or install the ldb tool.
In the particular case in question, the error message improved from a mere
```
Corruption: checksum mismatch
```
to
```
Corruption: checksum mismatch in file /path/to/db/engine-rocksdb/MANIFEST-000171
```
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9239
Reviewed By: jay-zhuang
Differential Revision: D33237742
Pulled By: riversand963
fbshipit-source-id: bd42559cfbf786a0a674d091671d1a2bf07bdd31
Summary:
Function `Version::UpdateFilesByCompactionPri()` is never called and not implemented.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/8724
Reviewed By: ajkr
Differential Revision: D30643943
Pulled By: riversand963
fbshipit-source-id: 174b2d9a2a42e286222909a035cc74a7b5602335
Summary:
Note: rebase on and merge after https://github.com/facebook/rocksdb/pull/9349, as part of https://github.com/facebook/rocksdb/pull/9342
**Context:**
https://github.com/facebook/rocksdb/pull/9073 charged the hash entries' memory in block cache with `CacheReservationHandle`. However, in the edge case where Ribbon Filter falls back to Bloom Filter and swaps its hash entries to the embedded bloom filter object, the handles associated with those entries are not swapped and thus not released as soon as those entries are cleared during Bloom Filter's finish process.
Although this is a minor issue since RocksDB internal calls `FilterBitsBuilder->Reset()` right after `FilterBitsBuilder->Finish()` on the main path, which releases all the cache reservation related to both the Ribbon Filter and its embedded Bloom Filter, it still worths this fix to avoid confusion.
**Summary:**
- Swapped the `CacheReservationHandle` associated with the hash entries on Ribbon Filter's fallback
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9345
Test Plan: - Added a unit test to verify the number of cache reservation after clearing hash entries, which failed before the change and now succeeds
Reviewed By: pdillinger
Differential Revision: D33377225
Pulled By: hx235
fbshipit-source-id: 7487f4c40dfb6ee7928232021f93ef2c5329cffa
Summary:
Closing https://github.com/facebook/rocksdb/issues/5006
Calling `DB::DestroyColumnFamilyHandle(column_family)` with `column_family` being the return value of
`DB::DefaultColumnFamily()` will return `Status::InvalidArgument()`.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9347
Test Plan: make check
Reviewed By: akankshamahajan15
Differential Revision: D33369675
Pulled By: riversand963
fbshipit-source-id: a8266a4daddf2b7a773c2dc7f3eb9a4adfb6b6dd
Summary:
Recently we added the ability to verify some prefix of operations are recovered (AKA no "hole" in the recovered data) (https://github.com/facebook/rocksdb/issues/8966). Besides testing unsynced data loss scenarios, it is also useful to test WAL disabled use cases, where unflushed writes are expected to be lost. Note RocksDB only offers the prefix-recovery guarantee to WAL-disabled use cases that use atomic flush, so crash test always enables atomic flush when WAL is disabled.
To verify WAL-disabled crash-recovery correctness globally, i.e., also in whitebox and blackbox transaction tests, it is possible but requires further changes. I added TODOs in db_crashtest.py.
Depends on https://github.com/facebook/rocksdb/issues/9305.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9338
Test Plan: Running all crash tests and many instances of blackbox. Sandcastle links are in Phabricator diff test plan.
Reviewed By: riversand963
Differential Revision: D33345333
Pulled By: ajkr
fbshipit-source-id: f56dd7d2e5a78d59301bf4fc3fedb980eb31e0ce
Summary:
The LastSequence field in the MANIFEST file is the baseline seqno for a recovered DB. Recovering WAL entries might cause the recovered DB's seqno to advance above this baseline, but the recovered DB will never use a smaller seqno.
Before this PR, we were writing the DB's seqno at the time of LogAndApply() as the LastSequence value. This works in the sense that it is a large enough baseline for the recovered DB that it'll never overwrite any records in existing SST files. At the same time, it's arbitrarily larger than what's needed. This behavior comes from LevelDB, where there was no tracking of largest seqno in an SST file.
Now we know the largest seqno of newly written SST files, so we can write an exact value in LastSequence that actually reflects the largest seqno in any file referred to by the MANIFEST. This is primarily useful for correctness testing with unsynced data loss, where the recovered DB's seqno needs to indicate what records were recovered.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9305
Test Plan:
- https://github.com/facebook/rocksdb/issues/9338 adds crash-recovery correctness testing coverage for WAL disabled use cases
- https://github.com/facebook/rocksdb/issues/9357 will extend that testing to cover file ingestion
- Added assertion at end of LogAndApply() for `VersionSet::descriptor_last_sequence_` consistency with files
- Manually tested upgrade/downgrade compatibility with a custom crash test that randomly picks between a `db_stress` built with and without this PR (for old code it must run with `-disable_wal=0`)
Reviewed By: riversand963
Differential Revision: D33182770
Pulled By: ajkr
fbshipit-source-id: 0bfafaf685f347cc8cb0e1d62e0186340a738f7d
Summary:
Allows the Env to have options (Configurable) and loads like other Customizable classes.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9293
Reviewed By: pdillinger, zhichao-cao
Differential Revision: D33181591
Pulled By: mrambacher
fbshipit-source-id: 55e823886c654d214eda9eedd45ccdc54dac14d7
Summary:
Fixes https://github.com/facebook/rocksdb/issues/9339
When writing SST file, the name, computed as `prefix_extractor->GetId()` will be written to the properties block.
When the SST is opened again in the future, `CreateFromString()` will take the name as argument and try
to create a prefix extractor object. Without this fix, the C API will pass a `Wrapper` pointer to the underlying
DB's `prefix_extractor`. `Wrapper::GetId()`, in this case, will be missing the prefix length component, causing a
prefix extractor of length 0 to be silently created and used.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9343
Test Plan:
```
make c_test
./c_test
```
Reviewed By: mrambacher
Differential Revision: D33355549
Pulled By: riversand963
fbshipit-source-id: c92c3acd8be262c3bff8794b4229e42b9ee31203
Summary:
Several improvements to SimulatedHybridFileSystem:
(1) Allow a mode where all I/Os to all files simulate HDD. This can be enabled in db_bench using -simulate_hdd
(2) Latency calculation is slightly more accurate
(3) Allow to simulate more than one HDD spindles.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9301
Test Plan: Run db_bench and observe the results are reasonable.
Reviewed By: jay-zhuang
Differential Revision: D33141662
fbshipit-source-id: b736e58c4ba910d06899cc9ccec79b628275f4fa
Summary:
Added new ObjectLibrary::Entry classes to replace/reduce the use of Regex. For simple factories that only do name matching, there are "StringEntry" and "AltStringEntry" classes. For classes that use some semblance of regular expressions, there is a PatternEntry class that can match a name and prefixes. There is also a class for Customizable::IndividualId format matches.
Added tests for the new derivative classes and got all unit tests to pass.
Resolves https://github.com/facebook/rocksdb/issues/9225.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9264
Reviewed By: pdillinger
Differential Revision: D33062001
Pulled By: mrambacher
fbshipit-source-id: c2d2143bd2d38bdf522705c8280c35381b135c03
Summary:
new blog post for Ribbon filter
Pull Request resolved: https://github.com/facebook/rocksdb/pull/8992
Test Plan: markdown render in GitHub, Pages on my fork
Reviewed By: jay-zhuang
Differential Revision: D33342496
Pulled By: pdillinger
fbshipit-source-id: a0a7c19100abdf8755f8a618eb4dead755dfddae
Summary:
This option causes trace records to be written in the serialized write thread. That way, the write records in the trace must follow the same order as writes that are logged to WAL and writes that are applied to the DB.
By default I left it disabled to match existing behavior. I enabled it in `db_stress`, though, as that use case requires order of write records in trace matches the order in WAL.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9334
Test Plan:
- See if below unsynced data loss crash test can run for 24h straight. It used to crash after a few hours when reaching an unlucky trace ordering.
```
DEBUG_LEVEL=0 TEST_TMPDIR=/dev/shm /usr/local/bin/python3 -u tools/db_crashtest.py blackbox --interval=10 --max_key=100000 --write_buffer_size=524288 --target_file_size_base=524288 --max_bytes_for_level_base=2097152 --value_size_mult=33 --sync_fault_injection=1 --test_batches_snapshots=0 --duration=86400
```
Reviewed By: zhichao-cao
Differential Revision: D33301990
Pulled By: ajkr
fbshipit-source-id: 82d97559727adb4462a7af69758449c8725b22d3
Summary:
As (https://github.com/facebook/rocksdb/issues/9210) discussed, the **full_history_ts_low** is a member of CompactRangeOptions currently, which means a CF's fullHistoryTsLow is advanced only when users submit a CompactRange request.
However, users may want to advance the fllHistoryTsLow without an immediate compact.
This merge make IncreaseFullHistoryTsLow to a public API so users can advance each CF's fullHistoryTsLow seperately.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9221
Reviewed By: akankshamahajan15
Differential Revision: D33201106
Pulled By: riversand963
fbshipit-source-id: 9cb1d013ba93260f72e16353e693ffee167b47ee
Summary:
The failure looked like this:
```
utilities/backupable/backupable_db_test.cc:3161: Failure
Value of: db_chroot_env_->FileExists(prev_manifest_path).IsNotFound()
Actual: false
Expected: true
```
The failure could be coerced consistently with the following patch:
```
diff --git a/db/db_impl/db_impl_compaction_flush.cc b/db/db_impl/db_impl_compaction_flush.cc
index 80410f671..637636791 100644
--- a/db/db_impl/db_impl_compaction_flush.cc
+++ b/db/db_impl/db_impl_compaction_flush.cc
@@ -2772,6 +2772,8 @@ void DBImpl::BackgroundCallFlush(Env::Priority thread_pri) {
if (job_context.HaveSomethingToClean() ||
job_context.HaveSomethingToDelete() || !log_buffer.IsEmpty()) {
mutex_.Unlock();
+ bg_cv_.SignalAll();
+ sleep(1);
TEST_SYNC_POINT("DBImpl::BackgroundCallFlush:FilesFound");
// Have to flush the info logs before bg_flush_scheduled_--
// because if bg_flush_scheduled_ becomes 0 and the lock is
```
The cause was a familiar problem, which is manual flush/compaction may
return before files they obsoleted are removed. The solution is just to
wait for "scheduled" work to complete, which includes all phases
including cleanup.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9327
Test Plan:
after this PR, even the above patch to coerce the bug cannot
cause the test to fail.
Reviewed By: riversand963
Differential Revision: D33252208
Pulled By: ajkr
fbshipit-source-id: 720a7eaca58c7247d221911fffe3d5e1dbf581e9
Summary:
locktree is a module providing Range Locking. It has a counter for
the number of times a lock acquisition request was blocked by an
existing conflicting lock and had to wait for it to be released.
Expose this counter in RangeLockManagerHandle::Counters::lock_wait_count.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9289
Reviewed By: jay-zhuang
Differential Revision: D33079182
Pulled By: riversand963
fbshipit-source-id: 25b1a362d9da247536ab5007bd15900b319f139e
Summary:
`db_stress` traces are used for tracking unsynced changes. For that purpose, we
only need to track writes and not reads. Currently `TraceOptions` only
supports excluding `Get()`s from the trace, so this PR only excludes
`Get()`s. In the future it would be good to exclude `MultiGet()`s and
iterator operations too.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9315
Test Plan:
- trace-heavy `db_stress` command elapsed time reduced 37%
Benchmark:
```
TEST_TMPDIR=/dev/shm /usr/bin/time ./db_stress -ops_per_thread=100000 -sync_fault_injection=1 -expected_values_dir=/dev/shm/dbstress_expected --clear_column_family_one_in=0
```
- replay-heavy `db_stress` command elapsed time reduced 38%
Setup:
```
TEST_TMPDIR=/dev/shm /usr/bin/time ./db_stress -ops_per_thread=100000000 -sync_fault_injection=1 -expected_values_dir=/dev/shm/dbstress_expected --clear_column_family_one_in=0 & sleep 120; pkill -9 db_stress
```
Benchmark:
```
TEST_TMPDIR=/dev/shm /usr/bin/time ./db_stress -ops_per_thread=1 -reopen=0 -expected_values_dir=/dev/shm/dbstress_expected --clear_column_family_one_in=0 --destroy_db_initially=0
```
Reviewed By: zhichao-cao
Differential Revision: D33229900
Pulled By: ajkr
fbshipit-source-id: 0e4251c674d236ddbc4548e9bbfdd608bf3cdc93
Summary:
* Added Docker build environment for RocksJava on s390x
* Cache alignment size for s390x was incorrectly calculated on gcc 6.4.0
* Tighter control over which installed version of Java is used is required - build now correctly adheres to `JAVA_HOME` if it is set
* Alpine build scripts should be used on Alpine (previously CentOS script worked by falling through to minimal gcc version)
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9321
Reviewed By: mrambacher
Differential Revision: D33259624
Pulled By: jay-zhuang
fbshipit-source-id: d791a5150581344925c3c3f9cbb9a3622d63b3b6
Summary:
There are some crashes we couldn't debug or repro and couldn't find a core dump. For ASAN the default is `disable_coredump=1` as the doc mentions core dumps can be 16TB+. However I've tried generating them for our `db_stress` commands and they've been in the 1.4-1.6GB range, which is fine. So we can try enabling it in CI.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9330
Test Plan:
- create a test job. (It's internal infra so I put the link in the Phabricator test plan only)
- ran the same command locally, `kill -6 $(pidof db_stress)`, verified core dump showed up
Reviewed By: jay-zhuang
Differential Revision: D33271841
Pulled By: ajkr
fbshipit-source-id: 93b853fa763d5708d078771960ba36854c4be55a
Summary:
Fixes a problem where the iterator for metadata was being treated as a non-user key when in fact it was a user key. This led to a problem where the property keys could not be searched for correctly.
The main exposure of this problem was that the HashIndexReader could not get the "prefixes" property correctly, resulting in the failure of retrieval/creation of the BlockPrefixIndex.
Added BlockBasedTableTest.SeekMetaBlocks test to validate this condition.
Fixing this condition exposed two other tests (SeekWithPrefixLongerThanKey, MultiGetPrefixFilter) that passed incorrectly previously and now failed. Updated those two tests to pass. Not sure if the tests are functionally correct/still appropriate, but made them pass...
Pull Request resolved: https://github.com/facebook/rocksdb/pull/8692
Reviewed By: riversand963
Differential Revision: D33119539
Pulled By: mrambacher
fbshipit-source-id: 658969fe9265f73dc184dab97cc3f4eaed2d881a
Summary:
We saw the below assertion failure in `error_handler_fs_test`:
```
db/error_handler_fs_test.cc:2471: Failure
Expected equality of these values:
listener->new_bg_error()
Which is: 16-byte object <00-00 00-00 00-00 00-00 00-00 00-00 00-00 00-00>
Status::Aborted()
Which is: 16-byte object <0A-00 00-00 60-61 00-00 00-00 00-00 00-00 00-00>
terminate called after throwing an instance of 'testing::internal::GoogleTestFailureException'
what(): db/error_handler_fs_test.cc:2471: Failure
Expected equality of these values:
listener->new_bg_error()
Which is: 16-byte object <00-00 00-00 00-00 00-00 00-00 00-00 00-00 00-00>
Status::Aborted()
Which is: 16-byte object <0A-00 00-00 60-61 00-00 00-00 00-00 00-00 00-00>
Received signal 6 (Aborted)
```
The problem was completing `OnErrorRecoveryCompleted()` would
wake up the main thread and allow it to proceed to that assertion. But
that assertion assumes `OnErrorRecoveryEnd()` has completed since
only `OnErrorRecoveryEnd()` affects `new_bg_error()`.
The fix is just to make `OnErrorRecoveryCompleted()` not wake up the
main thread, by means of not implementing it.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9325
Test Plan:
- ran `while TEST_TMPDIR=/dev/shm ./error_handler_fs_test ; do : ; done` for a while
- injected sleep between `OnErrorRecovery{Completed,End}()` callbacks, which guaranteed repro before this PR
Reviewed By: anand1976
Differential Revision: D33249200
Pulled By: ajkr
fbshipit-source-id: 1659ee183cd09f90d4dbd898f65103473fcf84a8
Summary:
I saw the following error when running crash test for a while with
unsynced data loss:
```
Error restoring historical expected values: Corruption: Corrupted trace file.
```
The trace file turned out to have an incomplete tail record. This is
normal considering blackbox kills `db_stress` while trace can be
ongoing.
In the case where the trace file is not otherwise corrupted, there
should be enough records already seen to sync up the expected state with
the recovered DB. This PR ignores any `Status::Corruption` the
`Replayer` returns when that happens.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9316
Reviewed By: jay-zhuang
Differential Revision: D33230579
Pulled By: ajkr
fbshipit-source-id: 9814af4e39e57f00d85be7404363211762f9b41b
Summary:
This fixes two bugs in the recently committed DB verification following
crash-recovery with unsynced data loss (https://github.com/facebook/rocksdb/issues/8966):
The first bug was in crash test runs involving mixed values for
`-test_batches_snapshots`. The problem was we were neither restoring
expected values nor enabling tracing when `-test_batches_snapshots=1`.
This caused a future `-test_batches_snapshots=0` run to not find enough
trace data to restore expected values. The fix is to restore expected
values at the start of `-test_batches_snapshots=1` runs, but still leave
tracing disabled as we do not need to track those KVs.
The second bug was in `db_stress` runs that restore the expected values
file and use compaction filter. The compaction filter was initialized to use
the pre-restore expected values, which would be `munmap()`'d during
`FileExpectedStateManager::Restore()`. Then compaction filter would run
into a segfault. The fix is just to reorder compaction filter init after expected
values restore.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9302
Test Plan:
- To verify the first problem, the below sequence used to fail; now it passes.
```
$ ./db_stress --db=./test-db/ --expected_values_dir=./test-db-expected/ --max_key=100000 --ops_per_thread=1000 --sync_fault_injection=1 --clear_column_family_one_in=0 --destroy_db_initially=0 -reopen=0 -test_batches_snapshots=0
$ ./db_stress --db=./test-db/ --expected_values_dir=./test-db-expected/ --max_key=100000 --ops_per_thread=1000 --sync_fault_injection=1 --clear_column_family_one_in=0 --destroy_db_initially=0 -reopen=0 -test_batches_snapshots=1
$ ./db_stress --db=./test-db/ --expected_values_dir=./test-db-expected/ --max_key=100000 --ops_per_thread=1000 --sync_fault_injection=1 --clear_column_family_one_in=0 --destroy_db_initially=0 -reopen=0 -test_batches_snapshots=0
```
- The second problem occurred rarely in the form of a SIGSEGV on a file that was `munmap()`d. I have not seen it after this PR though this doesn't prove much.
Reviewed By: jay-zhuang
Differential Revision: D33155283
Pulled By: ajkr
fbshipit-source-id: 66fd0f0edf34015a010c30015f14f104734e964e
Summary:
The `SharedState` constructor had an early return in case of
`-test_batches_snapshots=1`. This early return caused `num_bg_threads_`
to never be incremented. Consequently, the driver thread could cleanup
objects like the `SharedState` while BG threads were still running and
accessing it, leading to crash.
The fix is to move the logic for counting threads (both FG and BG) to
the place they are launched. That way we can be sure the counts are
consistent, at least for now.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9313
Test Plan:
below command used to fail, now it passes.
```
$ ./db_stress --db=./test-db/ --expected_values_dir=./test-db-expected/ --max_key=100000 --ops_per_thread=1000 --sync_fault_injection=1 --clear_column_family_one_in=0 --destroy_db_initially=0 -reopen=0 -test_batches_snapshots=1
```
Reviewed By: jay-zhuang
Differential Revision: D33198670
Pulled By: ajkr
fbshipit-source-id: 126592dc1eb31998bc8f82ffbf5a0d4eb8dec317
Summary:
in hope to get rockdb compiled with GCC-11 without warning
* util/bloom_test: init a variable before using it
to silence the GCC warning like
```
util/bloom_test.cc:1253:31: error: ‘<anonymous>’ may be used uninitialized [-Werror=maybe-uninitialized]
1253 | Slice key_slice{key_bytes, 8};
| ^
...
include/rocksdb/slice.h:41:3: note: by argument 2 of type ‘const char*’ to ‘rocksdb::Slice::Slice(const char*, size_t)’ declared here
41 | Slice(const char* d, size_t n) : data_(d), size_(n) {}
| ^~~~~
util/bloom_test.cc:1249:3: note: ‘<anonymous>’ declared here
1249 | };
| ^
cc1plus: all warnings being treated as errors
```
* cmake: add find_package(uring ...)
find liburing in a more consistent way. also it is the encouraged way for finding a library.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9286
Reviewed By: mrambacher
Differential Revision: D33165241
Pulled By: jay-zhuang
fbshipit-source-id: 9f3487e11b4e40fd8f1c97c8facb24a190e5ce31
Summary:
- Make MemoryAllocator and its implementations into a Customizable class.
- Added a "DefaultMemoryAllocator" which uses new and delete
- Added a "CountedMemoryAllocator" that counts the number of allocs and free
- Updated the existing tests to use these new allocators
- Changed the memkind allocator test into a generic test that can test the various allocators.
- Added tests for creating all of the allocators
- Added tests to verify/create the JemallocNodumpAllocator using its options.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/8980
Reviewed By: zhichao-cao
Differential Revision: D32990403
Pulled By: mrambacher
fbshipit-source-id: 6fdfe8218c10dd8dfef34344a08201be1fa95c76
Summary:
This change standardizes on a new 16-byte cache key format for
block cache (incl compressed and secondary) and persistent cache (but
not table cache and row cache).
The goal is a really fast cache key with practically ideal stability and
uniqueness properties without external dependencies (e.g. from FileSystem).
A fixed key size of 16 bytes should enable future optimizations to the
concurrent hash table for block cache, which is a heavy CPU user /
bottleneck, but there appears to be measurable performance improvement
even with no changes to LRUCache.
This change replaces a lot of disjointed and ugly code handling cache
keys with calls to a simple, clean new internal API (cache_key.h).
(Preserving the old cache key logic under an option would be very ugly
and likely negate the performance gain of the new approach. Complete
replacement carries some inherent risk, but I think that's acceptable
with sufficient analysis and testing.)
The scheme for encoding new cache keys is complicated but explained
in cache_key.cc.
Also: EndianSwapValue is moved to math.h to be next to other bit
operations. (Explains some new include "math.h".) ReverseBits operation
added and unit tests added to hash_test for both.
Fixes https://github.com/facebook/rocksdb/issues/7405 (presuming a root cause)
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9126
Test Plan:
### Basic correctness
Several tests needed updates to work with the new functionality, mostly
because we are no longer relying on filesystem for stable cache keys
so table builders & readers need more context info to agree on cache
keys. This functionality is so core, a huge number of existing tests
exercise the cache key functionality.
### Performance
Create db with
`TEST_TMPDIR=/dev/shm ./db_bench -bloom_bits=10 -benchmarks=fillrandom -num=3000000 -partition_index_and_filters`
And test performance with
`TEST_TMPDIR=/dev/shm ./db_bench -readonly -use_existing_db -bloom_bits=10 -benchmarks=readrandom -num=3000000 -duration=30 -cache_index_and_filter_blocks -cache_size=250000 -threads=4`
using DEBUG_LEVEL=0 and simultaneous before & after runs.
Before ops/sec, avg over 100 runs: 121924
After ops/sec, avg over 100 runs: 125385 (+2.8%)
### Collision probability
I have built a tool, ./cache_bench -stress_cache_key to broadly simulate host-wide cache activity
over many months, by making some pessimistic simplifying assumptions:
* Every generated file has a cache entry for every byte offset in the file (contiguous range of cache keys)
* All of every file is cached for its entire lifetime
We use a simple table with skewed address assignment and replacement on address collision
to simulate files coming & going, with quite a variance (super-Poisson) in ages. Some output
with `./cache_bench -stress_cache_key -sck_keep_bits=40`:
```
Total cache or DBs size: 32TiB Writing 925.926 MiB/s or 76.2939TiB/day
Multiply by 9.22337e+18 to correct for simulation losses (but still assume whole file cached)
```
These come from default settings of 2.5M files per day of 32 MB each, and
`-sck_keep_bits=40` means that to represent a single file, we are only keeping 40 bits of
the 128-bit cache key. With file size of 2\*\*25 contiguous keys (pessimistic), our simulation
is about 2\*\*(128-40-25) or about 9 billion billion times more prone to collision than reality.
More default assumptions, relatively pessimistic:
* 100 DBs in same process (doesn't matter much)
* Re-open DB in same process (new session ID related to old session ID) on average
every 100 files generated
* Restart process (all new session IDs unrelated to old) 24 times per day
After enough data, we get a result at the end:
```
(keep 40 bits) 17 collisions after 2 x 90 days, est 10.5882 days between (9.76592e+19 corrected)
```
If we believe the (pessimistic) simulation and the mathematical generalization, we would need to run a billion machines all for 97 billion days to expect a cache key collision. To help verify that our generalization ("corrected") is robust, we can make our simulation more precise with `-sck_keep_bits=41` and `42`, which takes more running time to get enough data:
```
(keep 41 bits) 16 collisions after 4 x 90 days, est 22.5 days between (1.03763e+20 corrected)
(keep 42 bits) 19 collisions after 10 x 90 days, est 47.3684 days between (1.09224e+20 corrected)
```
The generalized prediction still holds. With the `-sck_randomize` option, we can see that we are beating "random" cache keys (except offsets still non-randomized) by a modest amount (roughly 20x less collision prone than random), which should make us reasonably comfortable even in "degenerate" cases:
```
197 collisions after 1 x 90 days, est 0.456853 days between (4.21372e+18 corrected)
```
I've run other tests to validate other conditions behave as expected, never behaving "worse than random" unless we start chopping off structured data.
Reviewed By: zhichao-cao
Differential Revision: D33171746
Pulled By: pdillinger
fbshipit-source-id: f16a57e369ed37be5e7e33525ace848d0537c88f
Summary:
db_crashtest.py uses multiple CFs only when run without flag `--simple`.
The previous config set `-test_batches_snapshots=1` in that case for
blackbox mode. But `-test_batches_snapshots=1` cannot verify recovery
correctness, so it should not always be set for multi-CF blackbox tests.
We can instead randomly toggle it.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9303
Reviewed By: riversand963
Differential Revision: D33155229
Pulled By: ajkr
fbshipit-source-id: 4a6fdc4eddccc8ece664063baf6393ce1c5de6b7
Summary:
Existing multiGet() in java calls multi_get_helper() which then calls DB::std::vector MultiGet(). This doesn't take advantage of io_uring.
This change adds another JNI level method that runs a parallel code path using the DB::void MultiGet(), using ByteBuffers at the JNI level. We call it multiGetDirect(). In addition to using the io_uring path, this code internally returns pinned slices which we can copy out of into our direct byte buffers; this should reduce the overall number of copies in the code path to/from Java. Some jmh benchmark runs (100k keys, 1000 key multiGet) suggest that for value sizes > 1k, we see about a 20% performance improvement, although performance is slightly reduced for small value sizes, there's a little bit more overhead in the JNI methods.
Closes https://github.com/facebook/rocksdb/issues/8407
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9224
Reviewed By: mrambacher
Differential Revision: D32951754
Pulled By: jay-zhuang
fbshipit-source-id: 1f70df7334be2b6c42a9c8f92725f67c71631690
Summary:
When a previous run left behind historical state/trace files (implying it was run with --sync_fault_injection set), this PR uses them to restore the expected state according to the DB's recovered sequence number. That way, a tail of latest unsynced operations are permitted to be dropped, as is the case when data in page cache or certain `Env`s is lost. The point of the verification in this scenario is just to ensure there is no hole in the recovered data.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/8966
Test Plan:
- ran it a while, made sure it is restoring expected values using the historical state/trace files:
```
$ rm -rf ./tmp-db/ ./exp/ && mkdir -p ./tmp-db/ ./exp/ && while ./db_stress -compression_type=none -clear_column_family_one_in=0 -expected_values_dir=./exp -sync_fault_injection=1 -destroy_db_initially=0 -db=./tmp-db -max_key=1000000 -ops_per_thread=10000 -reopen=0 -threads=32 ; do : ; done
```
Reviewed By: pdillinger
Differential Revision: D31219445
Pulled By: ajkr
fbshipit-source-id: f0e1d51fe5b35465b00565c33331190ea38ba0ad
Summary:
SimulatedHybridFileSystem now takes a more thorough simualtion of an HDD:
1. cover writes too, not just read
2. Latency and throughput is now simulated as seek + read time, using a rate limiter
This implementation can be modified to simulate full HDD behavior, which is not yet done.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9259
Test Plan: Run db_bench and observe the desired behavior.
Reviewed By: jay-zhuang
Differential Revision: D32903039
fbshipit-source-id: a83f5d72143e114d5e75edf39d647bf0b71978e1
Summary:
Current db_stress does not cover complex read-write transactions. Therefore, this PR adds
coverage for emulated MyRocks-style transactions in `MultiOpsTxnsStressTest`. To achieve this, we need:
- Add a new operation type 'customops' so that we can add new complex groups of operations, e.g. transactions involving multiple read-write operations.
- Implement three read-write transactions and two read-only ones to emulate MyRocks-style transactions.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/8936
Test Plan:
```
make check
./db_stress -test_multi_ops_txns -use_txn -clear_column_family_one_in=0 -column_families=1 -writepercent=0 -delpercent=0 -delrangepercent=0 -customopspercent=60 -readpercent=20 -prefixpercent=0 -iterpercent=20 -reopen=0 -ops_per_thread=100000
```
Next step is to add more configurability and refine input generation and result reporting, which will done in separate follow-up PRs.
Reviewed By: zhichao-cao
Differential Revision: D31071795
Pulled By: riversand963
fbshipit-source-id: 50d7c828346ec643311336b904848a1588a37006
Jay Zhuang