[ https://issues.apache.org/jira/browse/CASSANDRA-5661?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13707684#comment-13707684 ]
Pavel Yaskevich commented on CASSANDRA-5661: -------------------------------------------- Thanks, it worked that time, I did a quick test on local machine with global cache backed by MultiwayPool. MultiwayPool average run: median 3.1, 95th 4.7, 99.9th 21.9 FileCacheService (+ LinkedTranferQueue) average run: median 2.4, 95th 3.0, 99.9th 17.1 It's the same setup that I used in my previous test with no writes and no compaction. I will try to experiment with ArrayBlockingQueue as we know upper bound on concurrency and update my FileCacheService patch with either of them (LTQ vs. ABQ) soon. > Discard pooled readers for cold data > ------------------------------------ > > Key: CASSANDRA-5661 > URL: https://issues.apache.org/jira/browse/CASSANDRA-5661 > Project: Cassandra > Issue Type: Bug > Components: Core > Affects Versions: 1.2.1 > Reporter: Jonathan Ellis > Assignee: Pavel Yaskevich > Fix For: 2.0 > > Attachments: CASSANDRA-5661-multiway-per-sstable.patch, > CASSANDRA-5661.patch, CASSANDRA-5661-v2-global-multiway-per-sstable.patch, > DominatorTree.png, Histogram.png > > > Reader pooling was introduced in CASSANDRA-4942 but pooled > RandomAccessReaders are never cleaned up until the SSTableReader is closed. > So memory use is "the worst case simultaneous RAR we had open for this file, > forever." > We should introduce a global limit on how much memory to use for RAR, and > evict old ones. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira