Summary:
Added an option for readrandom benchmark to run with tailing iterator instead of Get. Benefit of tailing iterator is that it doesn't require locking DB mutex on access.
I also have some results when running on my machine. The results highly depend on number of cache shards. With our current benchmark setting of 4 table cache shards and 6 block cache shards, I don't see much improvements of using tailing iterator. In that case, we're probably seeing cache mutex contention.
Here are the results for different number of shards
cache shards tailing iterator get
6 1.38M 1.16M
10 1.58M 1.15M
As soon as we get rid of cache mutex contention, we're seeing big improvements in using tailing iterator vs. ordinary get.
Test Plan: ran regression test
Reviewers: dhruba, haobo, ljin, kailiu, sding
Reviewed By: haobo
CC: tnovak
Differential Revision: https://reviews.facebook.net/D15867
Summary:
Previous I am too ambitious to hide every detail about table factory
to internal api. However, we cannot pass the compilatoin for external
users since we use table factory as the shared_ptr, which requires
the definition of table factory's destructor.
Test Plan: make check;
Reviewers: sdong, haobo
CC: leveldb
Differential Revision: https://reviews.facebook.net/D15861
Summary: RocksDB doesn't compile on 32-bit architecture apparently. This is attempt to fix some of 32-bit errors. They are reported here: https://gist.github.com/paxos/8789697
Test Plan: RocksDB still compiles on 64-bit :)
Reviewers: kailiu
Reviewed By: kailiu
CC: leveldb
Differential Revision: https://reviews.facebook.net/D15825
Summary:
Use super_version insider NewIterator to avoid Ref() each component
separately under mutex
The new added bench shows NewIterator QPS increases from 515K to 719K
No meaningful improvement for multiget I guess due to its relatively small
cost comparing to 90 keys fetch in the test.
Test Plan: unit test and db_bench
Reviewers: igor, sdong
Reviewed By: igor
CC: leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D15609
Summary: In PlainTable, use one single byte to represent 8 bytes of internal bytes, if seqID = 0 and it is value type (which should be common for bottom most files). It is to save 7 bytes for uncompressed cases.
Test Plan: make all check
Reviewers: haobo, dhruba, kailiu
Reviewed By: haobo
CC: igor, leveldb
Differential Revision: https://reviews.facebook.net/D15489
Summary:
VersionSet::next_file_number_ is always assumed to be strictly greater than VersionSet::log_number_. In our new recovery code, we artificially set log_number_ to be (log_number + 1), so that once we flush, we don't recover from the same log file again (this is important because of merge operator non-idempotence)
When we set VersionSet::log_number_ to (log_number + 1), we also have to mark that file number used, such that next_file_number_ is increased to a legal level. Otherwise, VersionSet might assert.
This has not be a problem so far because here's what happens:
1. assume next_file_number is 5, we're recovering log_number 10
2. in DBImpl::Recover() we call MarkFileNumberUsed with 10. This will set VersionSet::next_file_number_ to 11.
3. If there are some updates, we will call WriteTable0ForRecovery(), which will use file number 11 as a new table file and advance VersionSet::next_file_number_ to 12.
4. When we LogAndApply() with log_number 11, assertion is true: assert(11 <= 12);
However, this was a lucky occurrence. Even though this diff doesn't cause a bug, I think the issue is important to fix.
Test Plan: In column families I have different recovery logic and this code path asserted. When adding MarkFileNumberUsed(log_number + 1) assert is gone.
Reviewers: dhruba, kailiu
Reviewed By: kailiu
CC: leveldb
Differential Revision: https://reviews.facebook.net/D15783
Summary: I didn't figure out the reason why the feature of zeroing out earlier sequence ID is disabled in universal compaction. I do see bottommost_level is set correctly. It should simply work if we remove the constraint of universal compaction.
Test Plan: make all check
Reviewers: haobo, dhruba, kailiu, igor
Reviewed By: dhruba
CC: leveldb
Differential Revision: https://reviews.facebook.net/D15423
Summary:
Easy thing goes first. This patch moves arena to internal dir; based
on which, the coming patch will deal with memtable_rep.
Test Plan: make check
Reviewers: haobo, sdong, dhruba
CC: leveldb
Differential Revision: https://reviews.facebook.net/D15615
Summary: The default settings enable checksum verification on every read.
Test Plan: make check
Reviewers: haobo
Reviewed By: haobo
CC: leveldb
Differential Revision: https://reviews.facebook.net/D15591
Summary:
In DBImpl we keep track of some statistics internally and expose them via GetProperty(). This diff encapsulates all the internal statistics into a class InternalStatisics. Most of it is copy/paste.
Apart from cleaning up db_impl.cc, this diff is also necessary for Column families, since every column family should have its own CompactionStats, MakeRoomForWrite-stall stats, etc. It's much easier to keep track of it in every column family if it's nicely encapsulated in its own class.
Test Plan: make check
Reviewers: dhruba, kailiu, haobo, sdong, emayanke
Reviewed By: haobo
CC: leveldb
Differential Revision: https://reviews.facebook.net/D15273
Summary: as title
Test Plan: unit test
Reviewers: haobo, igor, sdong, kailiu, dhruba
Reviewed By: dhruba
CC: leveldb
Differential Revision: https://reviews.facebook.net/D15435
Summary:
I came across this while working on column families. CorruptionTest::RecoverWriteError threw a SIGSEG because the descriptor_log_->file() was nullptr. I'm not sure why it doesn't happen in master, but better safe than sorry.
@kailiu, can we get this in release, too?
Test Plan: make check
Reviewers: kailiu, dhruba, haobo
Reviewed By: haobo
CC: leveldb, kailiu
Differential Revision: https://reviews.facebook.net/D15513
Summary:
Blocks in the transaction log are a fixed size, but the last block in the transaction log file is usually a partial block. When a new record is added after the reader hit the end of the file, a new physical record will be appended to the last block. ReadPhysicalRecord can only read full blocks and assumes that the file position indicator is aligned to the start of a block. If the reader is forced to read further by simply clearing the EOF flag, ReadPhysicalRecord will read a full block starting from somewhere in the middle of a real block, causing it to lose alignment and to have a partial physical record at the end of the read buffer. This will result in length mismatches and checksum failures. When the log file is tailed for replication this will cause the log iterator to become invalid, necessitating the creation of a new iterator which will have to read the log file from scratch.
This diff fixes this issue by reading the remaining portion of the last block we read from. This is done when the reader is forced to read further (UnmarkEOF is called).
Test Plan:
- Added unit tests
- Stress test (with replication). Check dbdir/LOG file for corruptions.
- Test on test tier
Reviewers: emayanke, haobo, dhruba
Reviewed By: haobo
CC: vamsi, sheki, dhruba, kailiu, igor
Differential Revision: https://reviews.facebook.net/D15249
Summary:
@dhruba, I'm not sure where we need to sync the directory. I implemented the function in Env() and added the dir sync just after we close the newly created file in the builder.
Should I also add FsyncDir() to new files that get created by a compaction?
Test Plan: Confirmed that FsyncDir is returning Status::OK()
Reviewers: dhruba, haobo
Reviewed By: dhruba
CC: leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D14751
Summary: Converting from length prefixed buffer back to internal key costs some CPU but it is not necessary. In this patch, internal keys are pass though the functions so that we don't need to convert back to it.
Test Plan: make all check
Reviewers: haobo, kailiu
Reviewed By: kailiu
CC: igor, leveldb
Differential Revision: https://reviews.facebook.net/D15393
Summary: There is no reason to have functions NeedCompaction(), MaxCompactionScore() and MaxCompactionScoreLevel() in VersionSet, since they don't access any data in VersionSet.
Test Plan: make check
Reviewers: kailiu, haobo, sdong
Reviewed By: kailiu
CC: leveldb
Differential Revision: https://reviews.facebook.net/D15333
ReduceNumberOfLevels had segmentation fault in WriteSnapshot() since we
didn't change the number of levels in VersionSet (we consider them
immutable from now on). This fixes the problem.
Summary: By removing some includes form options.h and reply on forward declaration, we can more easily reason the dependencies.
Test Plan: make all check
Reviewers: kailiu, haobo, igor, dhruba
Reviewed By: kailiu
CC: leveldb
Differential Revision: https://reviews.facebook.net/D15411
Summary:
A lot of our code implicitly assumes number_levels to be static. ReduceNumberOfLevels() breaks that assumption. For example, after calling ReduceNumberOfLevels(), DBImpl::NumberLevels() will be different from VersionSet::NumberLevels(). This is dangerous. Thankfully, it's not in public headers and is only used from LDB cmd tool. LDB tool is only using it statically, i.e. it never calls it with running DB instance. With this diff, we make it explicitly static. This way, we can assume number_levels to be immutable and not break assumption that lot of our code is relying upon. LDB tool can still use the method.
Also, I removed the method from a separate file since it breaks filename completition. version_se<TAB> now completes to "version_set." instead of "version_set" (without the dot). I don't see a big reason that the function should be in a different file.
Test Plan: reduce_levels_test
Reviewers: dhruba, haobo, kailiu, sdong
Reviewed By: kailiu
CC: leveldb
Differential Revision: https://reviews.facebook.net/D15303
Summary:
MemTableListVersion is to MemTableList what Version is to VersionSet. I took almost the same ideas to develop MemTableListVersion. The reason is to have copying std::list done in background, while flushing, rather than in foreground (MultiGet() and NewIterator()) under a mutex! Also, whenever we copied MemTableList, we copied also some MemTableList metadata (flush_requested_, commit_in_progress_, etc.), which was wasteful.
This diff avoids std::list copy under a mutex in both MultiGet() and NewIterator(). I created a small database with some number of immutable memtables, and creating 100.000 iterators in a single-thread (!) decreased from {188739, 215703, 198028} to {154352, 164035, 159817}. A lot of the savings come from code under a mutex, so we should see much higher savings with multiple threads. Creating new iterator is very important to LogDevice team.
I also think this diff will make SuperVersion obsolete for performance reasons. I will try it in the next diff. SuperVersion gave us huge savings on Get() code path, but I think that most of the savings came from copying MemTableList under a mutex. If we had MemTableListVersion, we would never need to copy the entire object (like we still do in NewIterator() and MultiGet())
Test Plan: `make check` works. I will also do `make valgrind_check` before commit
Reviewers: dhruba, haobo, kailiu, sdong, emayanke, tnovak
Reviewed By: kailiu
CC: leveldb
Differential Revision: https://reviews.facebook.net/D15255
Summary:
Mixing index/filter blocks with data blocks resulted in some known
issues. To make sure in next release our users won't be affected,
we added a new option in BlockBasedTableFactory::TableOption to
conceal this functionality for now.
This patch also introduced a BlockBasedTableReader::OpenOptions,
which avoids the "infinite" growth of parameters in
BlockBasedTableReader::Open().
Test Plan: make check
Reviewers: haobo, sdong, igor, dhruba
Reviewed By: igor
CC: leveldb, tnovak
Differential Revision: https://reviews.facebook.net/D15327
Summary:
It looks like we might have some trouble when building the new release with 4.8, since fbcode is using glibc2.17-fb by default and we are using glibc2.17. It was reported by Benjamin Renard in our internal group.
This diff moves our fbcode build to use glibc2.17-fb by default. I got some linker errors when compiling, complaining that `google::SetUsageMessage()` was undefined. After deleting all offending lines, the compile was successful and everything works.
Test Plan:
Compiled
Ran ./db_bench ./db_stress ./db_repl_stress
Reviewers: kailiu
Reviewed By: kailiu
CC: leveldb
Differential Revision: https://reviews.facebook.net/D15405
Summary: Currently, compaction multipliers can overflow and cause unexpected behaviors. In this patch, we detect those overflows and use multiplier 1 for them.
Test Plan: make all check
Reviewers: dhruba, haobo, igor, kailiu
Reviewed By: kailiu
CC: leveldb
Differential Revision: https://reviews.facebook.net/D15321
Summary: as title
Test Plan:
make all check
What else tests shall I cover?
Reviewers: igor, haobo
CC:
Differential Revision: https://reviews.facebook.net/D15339
Summary:
This diff implements a special type of iterator that doesn't create a snapshot
(can be used to read newly inserted data) and is optimized for doing sequential
reads.
TailingIterator uses current superversion number to determine whether to
invalidate its internal iterators. If the version hasn't changed, it can often
avoid doing expensive seeks over immutable structures (sst files and immutable
memtables).
Test Plan:
* new unit tests
* running LD with this patch
Reviewers: igor, dhruba, haobo, sdong, kailiu
Reviewed By: sdong
CC: leveldb, lovro, march
Differential Revision: https://reviews.facebook.net/D15285
Summary:
In this diff I made some effort to reduce usage of friending. To do that, I had to expose Compaction::inputs_ through a method inputs(). Not sure if this is a good idea, there is a trade-off. I think it's less confusing than having lots of friends.
I also thought about other friendship relationships, but they are too much tangled at this point. Once you friend two classes, it's very hard to unfriend them :)
Test Plan: make check
Reviewers: haobo, kailiu, sdong, dhruba
Reviewed By: kailiu
CC: leveldb
Differential Revision: https://reviews.facebook.net/D15267
Summary:
This diff does two things:
* Rethinks how we call Recover() with read_only option. Before, we call it with pointer to memtable where we'd like to apply those changes to. This memtable is set in db_impl_readonly.cc and it's actually DBImpl::mem_. Why don't we just apply updates to mem_ right away? It seems more intuitive.
* Changes when we apply updates to manifest. Before, the process is to recover all the logs, flush it to sst files and then do one giant commit that atomically adds all recovered sst files and sets the next log number. This works good enough, but causes some small troubles for my column family approach, since I can't have one VersionEdit apply to more than single column family[1]. The change here is to commit the files recovered from logs right away. Here is the state of the world before the change:
1. Recover log 5, add new sst files to edit
2. Recover log 7, add new sst files to edit
3. Recover log 8, add new sst files to edit
4. Commit all added sst files to manifest and mark log files 5, 7 and 8 as recoverd (via SetLogNumber(9) function)
After the change, we'll do:
1. Recover log 5, commit the new sst files and set log 5 as recovered
2. Recover log 7, commit the new sst files and set log 7 as recovered
3. Recover log 8, commit the new sst files and set log 8 as recovered
The added (small) benefit is that if we fail after (2), the new recovery will only have to recover log 8. In previous case, we'll have to restart the recovery from the beginning. The bigger benefit will be to enable easier integration of multiple column families in Recovery code path.
[1] I'm happy to dicuss this decison, but I believe this is the cleanest way to go. It also makes backward compatibility much easier. We don't have a requirement of adding multiple column families atomically.
Test Plan: make check
Reviewers: dhruba, haobo, kailiu, sdong
Reviewed By: kailiu
CC: leveldb
Differential Revision: https://reviews.facebook.net/D15237
Summary: Commit "1304d8c8cefe66be1a3caa5e93413211ba2486f2" (Merge branch 'master' into performance) removes a line in performance branch by mistake. This patch fixes it.
Test Plan: make all check
Reviewers: haobo, kailiu, igor
Reviewed By: haobo
CC: leveldb
Differential Revision: https://reviews.facebook.net/D15297
Summary:
This moves the use of versions_ to before the mutex is unlocked
to avoid a possible race.
Task ID: #
Blame Rev:
Test Plan:
make check
Revert Plan:
Database Impact:
Memcache Impact:
Other Notes:
EImportant:
- begin *PUBLIC* platform impact section -
Bugzilla: #
- end platform impact -
Reviewers: haobo, dhruba
Reviewed By: dhruba
CC: leveldb
Differential Revision: https://reviews.facebook.net/D15279
Summary: I'm separating code-cleanup part of https://reviews.facebook.net/D14517. This will make D14517 easier to understand and this diff easier to review.
Test Plan: make check
Reviewers: haobo, kailiu, sdong, dhruba, tnovak
Reviewed By: tnovak
CC: leveldb
Differential Revision: https://reviews.facebook.net/D15099
Summary:
The SIGSEGV was introduced by https://reviews.facebook.net/D15171
I also fixed ExpandWhileOverlapping() which returned the failure by setting it's own stack variable to nullptr (!). This bug is present in 2.6 release, so I guess ExpandWhileOverlapping never fails :)
Test Plan: `make check`. Also MarkCallaghan confirmed it fixed the SIGSEGV he reported.
Reviewers: MarkCallaghan, kailiu, sdong, dhruba, haobo
Reviewed By: kailiu
CC: leveldb
Differential Revision: https://reviews.facebook.net/D15261
Summary:
This had a few bugs.
1) bottom and top were reversed. top is for the max value but the callers were passing the max
value to bottom. The result is that the max sleep is used when n >= bottom.
2) one of the callers passed values with type double and these values are frequently between
1.0 and 2.0 so rounding will do some bad things
3) sometimes the function returned 0 when there should be a stall
With this change and one other diff (out for review soon) there are slightly fewer stalls on one workload.
With the fix.
Stalls(secs): 160.166 level0_slowdown, 0.000 level0_numfiles, 0.000 memtable_compaction, 58.495 leveln_slowdown
Stalls(count): 910261 level0_slowdown, 0 level0_numfiles, 0 memtable_compaction, 54526 leveln_slowdown
Without the fix.
Stalls(secs): 172.227 level0_slowdown, 0.000 level0_numfiles, 0.000 memtable_compaction, 56.538 leveln_slowdown
Stalls(count): 160831 level0_slowdown, 0 level0_numfiles, 0 memtable_compaction, 52845 leveln_slowdown
Task ID: #
Blame Rev:
Test Plan:
run db_bench for --benchmarks=overwrite with IO-bound database
Revert Plan:
Database Impact:
Memcache Impact:
Other Notes:
EImportant:
- begin *PUBLIC* platform impact section -
Bugzilla: #
- end platform impact -
Reviewers: haobo
Reviewed By: haobo
CC: leveldb
Differential Revision: https://reviews.facebook.net/D15243
Summary:
This diff fixes 2 hacks:
* The callback function can modify the existing value inplace, if the merged value fits within the existing buffer size. But currently the existing buffer size is not being modified. Now the callback recieves a int* allowing the size to be modified. Since size is encoded as a varint in the internal key for memtable. It might happen that the entire value might have be copied to the new location if the new size varint is smaller than the existing size varint.
* The callback function has 3 functionalities
1. Modify existing buffer inplace, and update size correspondingly. Now to indicate that, Returns 1.
2. Generate a new buffer indicating merged value. Returns 2.
3. Fails to do either of above, based on whatever application logic. Returns 0.
Test Plan: Just make all for now. I'm adding another unit test to test each scenario.
Reviewers: dhruba, haobo
Reviewed By: haobo
CC: leveldb, sdong, kailiu, xinyaohu, sumeet, danguo
Differential Revision: https://reviews.facebook.net/D15195
