[ https://issues.apache.org/jira/browse/CASSANDRA-5541?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Jonathan Ellis updated CASSANDRA-5541: -------------------------------------- Attachment: 5541-v2.txt Ah, that makes sense. Looking at uses of RM.isEmpty, this is one of two places we expect to possibly create empty RM. Looks like the easiest fix would be to move the lazy-allocate into CLR -- this is the approach taken already in CassandraServer.createMutationList. v2 attached w/ that approach. > Reduce Empty Map allocations in RowMutation (modifications) > ----------------------------------------------------------- > > Key: CASSANDRA-5541 > URL: https://issues.apache.org/jira/browse/CASSANDRA-5541 > Project: Cassandra > Issue Type: Improvement > Components: Core > Affects Versions: 2.0 > Reporter: Dave Brosius > Assignee: Dave Brosius > Priority: Trivial > Fix For: 2.0 > > Attachments: 5541.txt, 5541-v2.txt > > > The constructor > public RowMutation(String table, ByteBuffer key) > was generation tons of empty HashMaps that were unused, on the order of > 160,000 of them in about a minute. > delay creating the map until you need it. -- 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