Find a file
Igor Canadi 7413306d94 Take a chance on a random file when choosing compaction
Summary:
When trying to compact entire database with SuggestCompactRange(), we'll first try the left-most files. This is pretty bad, because:
1) the left part of LSM tree will be overly compacted, but right part will not be touched
2) First compaction will pick up the left-most file. Second compaction will try to pick up next left-most, but this will not be possible, because there's a big chance that second's file range on N+1 level is already being compacted.

I observe both of those problems when running Mongo+RocksDB and trying to compact the DB to clean up tombstones. I'm unable to clean them up :(

This diff adds a bit of randomness into choosing a file. First, it chooses a file at random and tries to compact that one. This should solve both problems specified here.

Test Plan: make check

Reviewers: yhchiang, rven, sdong

Reviewed By: sdong

Subscribers: dhruba, leveldb

Differential Revision: https://reviews.facebook.net/D38379
2015-05-15 14:14:40 -07:00
arcanist_util Integrate Jenkins with Phabricator 2015-04-07 11:56:29 -07:00
build_tools Build for CYGWIN 2015-04-23 21:33:44 -07:00
coverage Fix coverage script 2014-11-03 14:53:00 -08:00
db Take a chance on a random file when choosing compaction 2015-05-15 14:14:40 -07:00
doc Remove seek compaction 2014-06-20 10:23:02 +02:00
examples Fix formatting 2015-02-09 09:53:30 -08:00
hdfs Remove unused parameter in CancelAllBackgroundWork 2015-03-16 21:07:54 -07:00
include Universal Compaction with multiple levels won't allocate up to output size 2015-05-13 14:15:46 -07:00
java Bugfix remove deprecated option use which was removed in previous commit 019ecd1932 2015-04-30 23:16:04 +01:00
port Build for CYGWIN 2015-04-23 21:33:44 -07:00
table Add more table properties to EventLogger 2015-05-12 15:53:55 -07:00
third-party Update COMMIT.md 2015-03-30 17:48:16 -07:00
tools Add Size-GB column to benchmark reports 2015-05-02 07:46:12 -07:00
util Make ThreadStatus::InterpretOperationProperties take const uint64_t* 2015-05-13 12:26:07 -07:00
utilities API to fetch from both a WriteBatchWithIndex and the db 2015-05-11 14:51:51 -07:00
.arcconfig Integrate Jenkins with Phabricator 2015-04-07 11:56:29 -07:00
.clang-format A script that automatically reformat affected lines 2014-01-14 12:21:24 -08:00
.gitignore run 'make check's rules (and even subtests) in parallel 2015-04-06 12:35:25 -07:00
.travis.yml Don't preinstall jemalloc in Travis 2015-04-24 18:43:07 -07:00
AUTHORS Add AUTHORS file. Fix #203 2014-09-29 10:52:18 -07:00
CONTRIBUTING.md facebook accounts are not required for CLA signers 2014-07-08 05:57:54 -04:00
HISTORY.md Task 6532943: Rocksdb - SetCapacity() can dynamically change cache capacity if feasible 2015-04-24 14:12:58 -07:00
INSTALL.md Optimize default compile to compilation platform by default 2014-12-15 11:29:41 +01:00
LICENSE Fix copyright year 2014-03-12 12:06:58 -07:00
Makefile Use version defined in Makefile in rocksdb_build_git_sha 2015-05-15 13:51:57 -07:00
PATENTS Update Patent Grant. 2015-04-13 10:33:43 +01:00
README.md Replaced "built on on earlier work" by "built on earlier work" in README.md 2014-09-17 01:16:17 -07:00
ROCKSDB_LITE.md RocksDBLite 2014-04-15 13:39:26 -07:00
src.mk Add more table properties to EventLogger 2015-05-12 15:53:55 -07:00
USERS.md Update USERS.md 2015-05-04 15:27:59 -07:00
Vagrantfile RocksDB on FreeBSD support 2015-02-26 15:19:17 -08:00

RocksDB: A Persistent Key-Value Store for Flash and RAM Storage

Build Status

RocksDB is developed and maintained by Facebook Database Engineering Team. It is built on earlier work on LevelDB by Sanjay Ghemawat (sanjay@google.com) and Jeff Dean (jeff@google.com)

This code is a library that forms the core building block for a fast key value server, especially suited for storing data on flash drives. It has a Log-Structured-Merge-Database (LSM) design with flexible tradeoffs between Write-Amplification-Factor (WAF), Read-Amplification-Factor (RAF) and Space-Amplification-Factor (SAF). It has multi-threaded compactions, making it specially suitable for storing multiple terabytes of data in a single database.

Start with example usage here: https://github.com/facebook/rocksdb/tree/master/examples

See the github wiki for more explanation.

The public interface is in include/. Callers should not include or rely on the details of any other header files in this package. Those internal APIs may be changed without warning.

Design discussions are conducted in https://www.facebook.com/groups/rocksdb.dev/