[ https://issues.apache.org/jira/browse/CASSANDRA-1608?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13091974#comment-13091974 ]
Jonathan Ellis commented on CASSANDRA-1608: ------------------------------------------- Committed with some changes: - Switched DataTracker registration to be eager instead of lazy -- otherwise sstables created before any compaction happened (during log replay for instance) would not be added to the manifest. Also added unregistration when a new Strategy is created on schema change. - added some debug logging to Manifest - moved the LevelDB classes from db.leveldb to db.compaction, so that I could add a "unqualified compaction strategy will be looked for in oac.db.compaction" rule - renamed LevelDB* to Leveled* - ran dos2unix on the intervaltree classes and reformatted - added code to make it not attempt to compact empty L0, which caused an assertion error - Added getUnleveledSSTables to CFSMBean Still to do: - CASSANDRA-3085 > Redesigned Compaction > --------------------- > > Key: CASSANDRA-1608 > URL: https://issues.apache.org/jira/browse/CASSANDRA-1608 > Project: Cassandra > Issue Type: Improvement > Components: Core > Reporter: Chris Goffinet > Assignee: Benjamin Coverston > Attachments: 1608-22082011.txt, 1608-v2.txt, 1608-v4.txt, 1608-v5.txt > > > After seeing the I/O issues in CASSANDRA-1470, I've been doing some more > thinking on this subject that I wanted to lay out. > I propose we redo the concept of how compaction works in Cassandra. At the > moment, compaction is kicked off based on a write access pattern, not read > access pattern. In most cases, you want the opposite. You want to be able to > track how well each SSTable is performing in the system. If we were to keep > statistics in-memory of each SSTable, prioritize them based on most accessed, > and bloom filter hit/miss ratios, we could intelligently group sstables that > are being read most often and schedule them for compaction. We could also > schedule lower priority maintenance on SSTable's not often accessed. > I also propose we limit the size of each SSTable to a fix sized, that gives > us the ability to better utilize our bloom filters in a predictable manner. > At the moment after a certain size, the bloom filters become less reliable. > This would also allow us to group data most accessed. Currently the size of > an SSTable can grow to a point where large portions of the data might not > actually be accessed as often. -- This message is automatically generated by JIRA. For more information on JIRA, see: http://www.atlassian.com/software/jira