[
https://issues.apache.org/jira/browse/CASSANDRA-13271?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15894204#comment-15894204
]
Robert Stupp commented on CASSANDRA-13271:
------------------------------------------
The change of the type of the parameter {{isMultiCell}} in
{{ListType.getInstace()}} from primitive {{boolean}} to {{Boolean}} doesn't
make sense.
||trunk|[branch|https://github.com/apache/cassandra/compare/trunk...snazy:13271-trunk]|[testall|http://cassci.datastax.com/view/Dev/view/snazy/job/snazy-13271-trunk-testall/lastSuccessfulBuild/]|[dtest|http://cassci.datastax.com/view/Dev/view/snazy/job/snazy-13271-trunk-dtest/lastSuccessfulBuild/]
When the above issue is resolved, I think we can commit this one.
> Reduce lock contention on instance factories of ListType and SetType
> --------------------------------------------------------------------
>
> Key: CASSANDRA-13271
> URL: https://issues.apache.org/jira/browse/CASSANDRA-13271
> Project: Cassandra
> Issue Type: Improvement
> Components: Core
> Reporter: vincent royer
> Priority: Minor
> Labels: performance
> Fix For: 4.x
>
> Attachments: 0001-CASSANDRA-13271-computeIfAbsent.patch,
> 0001-CASSANDRA-13271-singleton-factory-concurrency-opimiz.patch
>
>
> By doing some performance tests, i noticed that getInstance() in
> org.apache.cassandra.db.marshal.ListType and SetType could suffer from lock
> contention on the singleton factory getInstance(). Here is a proposal to
> reduce lock contention by using a ConcurrentMap and the putIfAbsent method
> rather than a synchronized method.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
