Commit Graph

176 Commits

Author SHA1 Message Date
Lei Jin
0d755fff14 cache friendly blocked bloomfilter
Summary:
By constraining the probes within cache line(s), we can improve the
cache miss rate thus performance. This probably only makes sense for
in-memory workload so defaults the option to off.

Numbers and comparision can be found in wiki:
https://our.intern.facebook.com/intern/wiki/index.php/Ljin/rocksdb_perf/2014_03_17#Bloom_Filter_Study

Test Plan: benchmarked this change substantially. Will run make all check as well

Reviewers: haobo, igor, dhruba, sdong, yhchiang

Reviewed By: haobo

CC: leveldb

Differential Revision: https://reviews.facebook.net/D17133
2014-03-28 09:21:20 -07:00
Kai Liu
69f6cf431d Fix two bugs in talbe format
Summary:
Previous code had two bugs:

* didn't initialize the table_magic_number_ explicitly -- as a
  result a random junk number is stored for table_magic_number_, making
  HasInitializedMagicNumber() always return true.
* if condition is inconrrect in set_table_magic_number(), and the return value is not checked.
  I replace if-else by a stronger requirement enforced by assert().

Test Plan:
Previous sst_dump failed to work.
After the fix, things back to normal.

Reviewers: yhchiang

CC: haobo, sdong, igor, dhruba, leveldb

Differential Revision: https://reviews.facebook.net/D17055
2014-03-19 16:18:33 -07:00
Lei Jin
8d007b4aaf Consolidate SliceTransform object ownership
Summary:
(1) Fix SanitizeOptions() to also check HashLinkList. The current
dynamic case just happens to work because the 2 classes have the same
layout.
(2) Do not delete SliceTransform object in HashSkipListFactory and
HashLinkListFactory destructor. Reason: SanitizeOptions() enforces
prefix_extractor and SliceTransform to be the same object when
Hash**Factory is used. This makes the behavior strange: when
Hash**Factory is used, prefix_extractor will be released by RocksDB. If
other memtable factory is used, prefix_extractor should be released by
user.

Test Plan: db_bench && make asan_check

Reviewers: haobo, igor, sdong

Reviewed By: igor

CC: leveldb, dhruba

Differential Revision: https://reviews.facebook.net/D16587
2014-03-10 12:56:46 -07:00
Kai Liu
566f18e6ad More precise calculation of sub_index_size
Summary:
Previous we did rough estimation of subindex size, which in worst case may result in array reallocation.
This patch aims to get the exact size and avoid any reallocation.

Test Plan: make all check

Reviewers: sdong, dhruba, haobo

Reviewed By: haobo

CC: leveldb

Differential Revision: https://reviews.facebook.net/D16125
2014-03-06 17:30:46 -08:00
kailiu
a01bda0997 Fix a buggy assert
Summary: The assert was pointless since if if prefix is the same as the whole key, assertion will surely fail. Reason behind is when performing the internal key comparison, if user keys are the same, *key with smaller transaction id* wins.

Test Plan: make -j32 && make check

Reviewers: sdong, dhruba, haobo

CC: leveldb

Differential Revision: https://reviews.facebook.net/D16551
2014-03-04 18:08:23 -08:00
kailiu
a1d56e73ec Uncomment the unit tests in table test 2014-03-03 22:19:49 -08:00
kailiu
906f3dca72 Add a hash-index component for block
Summary:
this is the key component extracted from diff: https://reviews.facebook.net/D14271
I separate it to a dedicated patch to make the review easier.

Test Plan: added a unit test and passed it.

Reviewers: haobo, sdong, dhruba

CC: leveldb

Differential Revision: https://reviews.facebook.net/D16245
2014-03-03 21:11:49 -08:00
sdong
f0ee2356af Fix issue with iterator operations in this order: Prev(), Seek(), Prev()
Summary:
Due to a bad merge of D14163 and D14001 before checking in D14001, "direction_ = kForward;" in MergeIterator::Seek() was deleted my mistake (in commit b135d01e7b ). It will generate wrong results or assert failure after the sequence of Prev() (or SeekToLast()), Seek() and Prev().

Fix it

Test Plan: make all check

Reviewers: igor, haobo, dhruba

Reviewed By: igor

CC: yhchiang, i.am.jin.lei, ljin, leveldb

Differential Revision: https://reviews.facebook.net/D16527
2014-03-03 17:50:27 -08:00
kailiu
1aeafeccac Make the Create() function comform the convention
Summary:

Moved "Return multiple values" a more conventional way.
2014-03-01 23:43:03 -08:00
Kai Liu
16d4e45c12 Fix the memory leak in table index
Summary:

BinarySearchIndex didn't use unique_ptr to guard the block object nor
delete it in destructor, leading to valgrind failure for "definite
memory leak".

Test Plan:
re-ran the failed valgrind test cases
2014-03-01 11:50:35 -08:00
Kai Liu
ff151132b3 Fix the unit test failure in devbox
Summary:
My last diff was developed in MacOS but in devserver environment error occurs.

I dug into the problem and found the way we calcuate approximate data size is pretty out-of-date. We can use table properties to get more accurate results.

Test Plan: ran ./table_test and passed

Reviewers: igor, dhruba, haobo, sdong

CC: leveldb

Differential Revision: https://reviews.facebook.net/D16509
2014-02-28 20:40:05 -08:00
kailiu
74939a9e13 Make the block-based table's index pluggable
Summary:
This patch introduced a new table options that allows us to make
block-based table's index pluggable.

To support that new features:

* Code has been refacotred to be more flexible and supports this option well.
* More documentation is added for the existing obsecure functionalities.
* Big surgeon on DataBlockReader(), where the logic was really convoluted.
* Other small code cleanups.

The pluggablility will mostly affect development of internal modules
and won't change frequently, as a result I intentionally avoid
heavy-weight patterns (like factory) and try to make it simple.

Test Plan: make all check

Reviewers: haobo, sdong

Reviewed By: sdong

CC: leveldb

Differential Revision: https://reviews.facebook.net/D16395
2014-02-28 18:19:07 -08:00
kailiu
bf86af5174 Remove the terrible hack in for flush_block_policy_factory
Summary:
Previous code is too convoluted and I must be drunk for letting
such code to be written without a second thought.

Thanks to the discussion with @sdong, I added the `Options` when
generating the flusher, thus avoiding the tricks.

Just FYI: I resisted to add Options in flush_block_policy.h since I
wanted to avoid cyclic dependencies: FlushBlockPolicy dpends on Options
and Options also depends FlushBlockPolicy... While I appreciate my
effort to prevent it, the old design turns out creating more troubles than
it tried to avoid.

Test Plan: ran ./table_test

Reviewers: sdong

Reviewed By: sdong

CC: sdong, leveldb

Differential Revision: https://reviews.facebook.net/D16503
2014-02-28 16:39:27 -08:00
kailiu
444cafc28c Fix inconsistent code format
Summary:
Found some function follows camel style. When naming funciton, we have two styles:

Trivially expose internal data in readonly mode: `all_lower_case()`
Regular function: `CapitalizeFirstLetter()`

I renames these functions.

Test Plan: make -j32

Reviewers: haobo, sdong, dhruba, igor

CC: leveldb

Differential Revision: https://reviews.facebook.net/D16383
2014-02-26 18:56:39 -08:00
sdong
a04dbf6e49 PlainTable::Next() should pass the error message from ReadKey()
Summary:
PlainTable::Next() should pass the error message from ReadKey(). Now it would return a wrong error message.
Also improve the messages of status when failing to read

Test Plan: make all check

Reviewers: ljin, kailiu, haobo

Reviewed By: kailiu

CC: leveldb

Differential Revision: https://reviews.facebook.net/D16365
2014-02-26 15:12:44 -08:00
Kai Liu
2b205b35d8 Disable putting filter block to block cache
Summary: This bug caused server crash issues because the filter block is too big and kept purging out of cache.

Test Plan: Wrote a new unit tests to make sure it works.

Reviewers: dhruba, haobo, igor, sdong

Reviewed By: haobo

CC: leveldb

Differential Revision: https://reviews.facebook.net/D16221
2014-02-19 15:38:57 -08:00
kailiu
63690625cd Expose the table properties to application
Summary: Provide a public API for users to access the table properties for each SSTable.

Test Plan: Added a unit tests to test the function correctness under differnet conditions.

Reviewers: haobo, dhruba, sdong

Reviewed By: haobo

CC: leveldb

Differential Revision: https://reviews.facebook.net/D16083
2014-02-13 16:28:21 -08:00
Kai Liu
b2e7ee8b41 Followup code refactor on plain table
Summary:
Fixed most comments in https://reviews.facebook.net/D15429.
Still have some remaining comments left.

Test Plan: make all check

Reviewers: sdong, haobo

Reviewed By: haobo

CC: leveldb

Differential Revision: https://reviews.facebook.net/D15885
2014-02-13 15:27:59 -08:00
Kai Liu
59cffe02c4 Benchmark table reader wiht nanoseconds
Summary: nanosecnods gave us better view of the performance, especially when some operations are fast so that micro seconds may only reveal less informative results.

Test Plan:
sample output:

    ./table_reader_bench --plain_table --time_unit=nanosecond
    =======================================================================================================
    InMemoryTableSimpleBenchmark:           PlainTable   num_key1:   4096   num_key2:   512   non_empty
    =======================================================================================================
    Histogram (unit: nanosecond):
    Count: 6291456  Average: 475.3867  StdDev: 556.05
    Min: 135.0000  Median: 400.1817  Max: 33370.0000
    Percentiles: P50: 400.18 P75: 530.02 P99: 887.73 P99.9: 8843.26 P99.99: 9941.21
    ------------------------------------------------------
    [     120,     140 )        2   0.000%   0.000%
    [     140,     160 )      452   0.007%   0.007%
    [     160,     180 )    13683   0.217%   0.225%
    [     180,     200 )    54353   0.864%   1.089%
    [     200,     250 )   101004   1.605%   2.694%
    [     250,     300 )   729791  11.600%  14.294% ##
    [     300,     350 )   616070   9.792%  24.086% ##
    [     350,     400 )  1628021  25.877%  49.963% #####
    [     400,     450 )   647220  10.287%  60.250% ##
    [     450,     500 )   577206   9.174%  69.424% ##
    [     500,     600 )  1168585  18.574%  87.999% ####
    [     600,     700 )   506875   8.057%  96.055% ##
    [     700,     800 )   147878   2.350%  98.406%
    [     800,     900 )    42633   0.678%  99.083%
    [     900,    1000 )    16304   0.259%  99.342%
    [    1000,    1200 )     7811   0.124%  99.466%
    [    1200,    1400 )     1453   0.023%  99.490%
    [    1400,    1600 )      307   0.005%  99.494%
    [    1600,    1800 )       81   0.001%  99.496%
    [    1800,    2000 )       18   0.000%  99.496%
    [    2000,    2500 )        8   0.000%  99.496%
    [    2500,    3000 )        6   0.000%  99.496%
    [    3500,    4000 )        3   0.000%  99.496%
    [    4000,    4500 )      116   0.002%  99.498%
    [    4500,    5000 )     1144   0.018%  99.516%
    [    5000,    6000 )     1087   0.017%  99.534%
    [    6000,    7000 )     2403   0.038%  99.572%
    [    7000,    8000 )     9840   0.156%  99.728%
    [    8000,    9000 )    12820   0.204%  99.932%
    [    9000,   10000 )     3881   0.062%  99.994%
    [   10000,   12000 )      135   0.002%  99.996%
    [   12000,   14000 )      159   0.003%  99.998%
    [   14000,   16000 )       58   0.001%  99.999%
    [   16000,   18000 )       30   0.000% 100.000%
    [   18000,   20000 )       14   0.000% 100.000%
    [   20000,   25000 )        2   0.000% 100.000%
    [   25000,   30000 )        2   0.000% 100.000%
    [   30000,   35000 )        1   0.000% 100.000%

Reviewers: haobo, dhruba, sdong

CC: leveldb

Differential Revision: https://reviews.facebook.net/D16113
2014-02-13 13:57:36 -08:00
sdong
b5140a0361 Fix table_reader_bench and add it to "make"
Summary: Fix table_reader_bench after some interface changes. Add it to make to avoid future breaking

Test Plan: make table_reader_bench and run it with different options.

Reviewers: kailiu, haobo

Reviewed By: haobo

CC: igor, leveldb

Differential Revision: https://reviews.facebook.net/D16107
2014-02-12 18:31:02 -08:00
Siying Dong
f3ae3d07cc Add more black-box tests for PlainTable and explicitly support total order mode
Summary:
1. Add some more implementation-aware tests for PlainTable
2. move from a hard-coded one index per 16 rows in one prefix to a configurable number. Also, make hash table ratio = 0  means binary search only. Also fixes some divide 0 risks.
3. Explicitly support total order (only use binary search)
4. some code cleaning up.

Test Plan: make all check

Reviewers: haobo, kailiu

Reviewed By: haobo

CC: leveldb

Differential Revision: https://reviews.facebook.net/D16023
2014-02-12 17:37:22 -08:00
kailiu
e6b3e3b4db Support prefix seek in UserCollectedProperties
Summary: We'll need the prefix seek support for property aggregation.

Test Plan: make all check

Reviewers: haobo, sdong, dhruba

Reviewed By: haobo

CC: leveldb

Differential Revision: https://reviews.facebook.net/D15963
2014-02-12 13:14:59 -08:00
kailiu
aa734ce9ab Fix a member variables initialization order issue
Summary:
In MacOS, I got issue with `Footer`'s default constructor, which initialized the magic number with some random number instead of 0.
With investigation, I found we forgot to make the kInvalidTableMagicNumber to be static. As a result, kInvalidTableMagicNumber was assgined to `table_magic_number_` before it is initialized (which will be populated with random number).

Test Plan: passed current unit tests; also passed the unit tests for the incoming diff which used the default footer.

Reviewers: yhchiang

CC: leveldb

Differential Revision: https://reviews.facebook.net/D16077
2014-02-11 14:16:46 -08:00
kailiu
745c181e20 Quick fix for table_test failure
Summary:
* Fixed the compression state array size bug.
* Temporarily disable running `DoCompressionTest()` against bzip, which will fail the test.

Test Plan: make && ./table_test

Reviewers: igor

CC: leveldb

Differential Revision: https://reviews.facebook.net/D16065
2014-02-10 17:05:14 -08:00
Igor Canadi
8e634d3ea4 Merge pull request #74 from alberts/lz4
Support for LZ4 compression.
2014-02-10 15:46:56 -08:00
Albert Strasheim
df2f92214a Support for LZ4 compression. 2014-02-08 14:15:51 -08:00
Kai Liu
9a270f3f6d Fix the valgrind error in table test. 2014-02-07 22:43:58 -08:00
Kai Liu
b8ea5e36b3 Fix incompatible compilation in Linux server 2014-02-07 19:47:48 -08:00
kailiu
161ab42a8a Make table properties shareable
Summary:
We are going to expose properties of all tables to end users through "some" db interface.
However, current design doesn't naturally fit for this need, which is because:

1. If a table presents in table cache, we cannot simply return the reference to its table properties, because the table may be destroy after compaction (and we don't want to hold the ref of the version).
2. Copy table properties is OK, but it's slow.

Thus in this diff, I change the table reader's interface to return a shared pointer (for const table properties), instead a const refernce.

Test Plan: `make check` passed

Reviewers: haobo, sdong, dhruba

Reviewed By: haobo

CC: leveldb

Differential Revision: https://reviews.facebook.net/D15999
2014-02-07 19:26:49 -08:00
Yueh-Hsuan Chiang
3ce8d9a988 Add support for plain table format to sst_dump.
Summary:
This diff enables the command line tool `sst_dump` to work for sst files
under plain table format.  Changes include:
  * In tools/sst_dump.cc:
    - add support for plain table format
    - display prefix_extractor information when --show_properties is on
  * In table/format.cc
    - Now the table magic number of a Footer can be later initialized
      via ReadFooterFromFile().
  * In table/meta_bocks:
    - add function ReadTableMagicNumber() that reads the magic number of
      the specified file.

Minor fixes:
 - remove a duplicate #include in table/table_test.cc
 - fix a commentary typo in include/rocksdb/memtablerep.h
 - fix lint errors.

Test Plan:
Runs sst_dump with both block-based and plain-table format files with
different arguments, specifically those with --show-properties and --from.

* sample output:
  https://reviews.facebook.net/P261

Reviewers: kailiu, sdong, xjin

CC: leveldb

Differential Revision: https://reviews.facebook.net/D15903
2014-02-07 11:15:00 -08:00
kailiu
84f8185fc0 Merge branch 'master' into performance
Conflicts:
	HISTORY.md
	db/db_impl.cc
	db/memtable.cc
2014-02-05 21:21:00 -08:00
kailiu
d43ebd8c65 Put table factory back to public api
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
2014-02-03 19:51:20 -08:00
Igor Canadi
2966d764cd Fix some 32-bit compile errors
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
2014-02-03 13:48:30 -08:00
Siying Dong
d169b67680 [Performance Branch] PlainTable to encode rows with seqID 0, value type using 1 internal byte.
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
2014-02-03 12:19:30 -08:00
kailiu
4f6cb17bdb First phase API clean up
Summary:
Addressed all the issues in https://reviews.facebook.net/D15447.
Now most table-related modules are hidden from user land.

Test Plan: make check

Reviewers: sdong, haobo, dhruba

CC: leveldb

Differential Revision: https://reviews.facebook.net/D15525
2014-02-03 00:30:43 -08:00
kailiu
4e0298f23c Clean up arena API
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
2014-01-30 22:10:10 -08:00
kailiu
a5e220f5ef Merge branch 'master' into performance
Conflicts:
	Makefile
	db/db_impl.cc
	db/db_test.cc
	db/memtable_list.cc
	db/memtable_list.h
	table/block_based_table_reader.cc
	table/table_test.cc
	util/cache.cc
	util/coding.cc
2014-01-28 10:35:55 -08:00
Kai Liu
4b51dffcf8 Some refactorings on plain table
Summary:
Plain table has been working well and this is just a nit-picking patch,
which is generated during my coding reading. No real functional changes.
only some changes regarding:

* Improve some comments from the perspective a "new" code reader.
* Change some magic number to constant, which can help us to parameterize them
  in the future.
* Did some style, naming, C++ convention changes.
* Fix warnings from new "arc lint"

Test Plan: make check

Reviewers: sdong, haobo

CC: leveldb

Differential Revision: https://reviews.facebook.net/D15429
2014-01-24 21:28:10 -08:00
Kai Liu
0ab766132b Re-org the table tests
Summary:
We'll divide the table tests into 3 buckets, plain table test, block-based table test and general table feature test.
This diff does no real change and only does the rename and reorg.

Test Plan: run table_test

Reviewers: sdong, haobo, igor, dhruba

CC: leveldb

Differential Revision: https://reviews.facebook.net/D15417
2014-01-24 12:46:17 -08:00
Kai Liu
7d991be400 Some small refactorings on table_test
Summary:

Just revise some hard-to-read or unnecessarily verbose code.

Test Plan:

make check
2014-01-24 11:10:32 -08:00
kailiu
66dc033af3 Temporarily disable caching index/filter blocks
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
2014-01-24 10:57:15 -08:00
Kai Liu
054c5dda8c Merge branch 'master' into performance
Conflicts:
	db/db_impl.cc
	db/db_test.cc
	db/memtable.cc
	db/version_set.cc
	include/rocksdb/statistics.h
	util/statistics_imp.h
2014-01-23 16:32:49 -08:00
Kai Liu
ef602f6275 Misc cleanup on performance branch
Summary:

Did some trivial stuffs:

* Add more comments;
* fix compiler's warning messages (uninitialized variables).
* etc

Test Plan:

make check
2014-01-17 14:26:29 -08:00
Igor Canadi
83681bf9ef Statistics code cleanup
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
2014-01-17 12:46:06 -08:00
kailiu
1304d8c8ce Merge branch 'master' into performance
Conflicts:
	Makefile
	db/db_impl.cc
	db/db_impl.h
	db/db_test.cc
	db/memtable.cc
	db/memtable.h
	db/version_edit.h
	db/version_set.cc
	include/rocksdb/options.h
	util/hash_skiplist_rep.cc
	util/options.cc
2014-01-15 23:12:31 -08:00
Igor Canadi
7f3e417f59 Fix memtable construction in tests 2014-01-14 15:36:12 -08:00
Kai Liu
8c4eb71b5d Fix one more valgrind error in table_test 2014-01-03 18:27:33 -08:00
Kai Liu
5e7d5629c7 Fix the valgrind issues 2014-01-03 11:48:31 -08:00
Siying Dong
abaf26266d [RocksDB] [Performance Branch] Some Changes to PlainTable format
Summary:
Some changes to PlainTable format:
(1) support variable key length
(2) use user defined slice transformer to extract prefixes
(3) Run some test cases against PlainTable in db_test and table_test

Test Plan: test db_test

Reviewers: haobo, kailiu

CC: dhruba, igor, leveldb, nkg-

Differential Revision: https://reviews.facebook.net/D14457
2013-12-20 12:08:35 -08:00
Siying Dong
28c24de8be [RocksDB Peformance Branch] A bug in PlainTable format
Summary: A bug to fix. IT's already fixed in D14457, but want to check it in sooner to unblock tests

Test Plan: plain_table_db_test

Reviewers: nkg-, haobo

Reviewed By: nkg-

CC: kailiu, leveldb

Differential Revision: https://reviews.facebook.net/D14673
2013-12-13 21:51:16 -08:00