[ https://issues.apache.org/jira/browse/CASSANDRA-8546?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14260811#comment-14260811 ]
Philip Thompson edited comment on CASSANDRA-8546 at 12/30/14 5:17 AM: ---------------------------------------------------------------------- 2.0 is accepting bug fixes only at this point. This patch will be included in a 2.1 release, if accepted. This patch seems targeted towards 2.0, please rebase it for 2.1. was (Author: philipthompson): 2.0 is accepting bug fixes only at this point. This patch will be included in a 2.1 release, if accepted. > RangeTombstoneList becoming bottleneck on tombstone heavy tasks > --------------------------------------------------------------- > > Key: CASSANDRA-8546 > URL: https://issues.apache.org/jira/browse/CASSANDRA-8546 > Project: Cassandra > Issue Type: Improvement > Components: Core > Environment: 2.0.11 / 2.1 > Reporter: Dominic Letz > Fix For: 2.1.3 > > Attachments: cassandra-2.0.11-8546.txt, tombstone_test.tgz > > > I would like to propose a change of the data structure used in the > RangeTombstoneList to store and insert tombstone ranges to something with at > least O(log N) insert in the middle and at near O(1) and start AND end. Here > is why: > When having tombstone heavy work-loads the current implementation of > RangeTombstoneList becomes a bottleneck with slice queries. > Scanning the number of tombstones up to the default maximum (100k) can take > up to 3 minutes of how addInternal() scales on insertion of middle and start > elements. > The attached test shows that with 50k deletes from both sides of a range. > INSERT 1...110000 > flush() > DELETE 1...50000 > DELETE 110000...60000 > While one direction performs ok (~400ms on my notebook): > {code} > SELECT * FROM timeseries WHERE name = 'a' ORDER BY timestamp DESC LIMIT 1 > {code} > The other direction underperforms (~7seconds on my notebook) > {code} > SELECT * FROM timeseries WHERE name = 'a' ORDER BY timestamp ASC LIMIT 1 > {code} -- This message was sent by Atlassian JIRA (v6.3.4#6332)