[
https://issues.apache.org/jira/browse/HBASE-11537?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14067201#comment-14067201
]
Mike Drob commented on HBASE-11537:
-----------------------------------
[~jmhsieh] - Thanks for looking. I agree that the existing implementation did
not have a race condition, but I'm worried that there were already unguarded
get/remove calls and a change in behaviour in surrounding code might expose
something.
> Avoid synchronization on instances of ConcurrentMap
> ---------------------------------------------------
>
> Key: HBASE-11537
> URL: https://issues.apache.org/jira/browse/HBASE-11537
> Project: HBase
> Issue Type: Bug
> Reporter: Mike Drob
> Assignee: Mike Drob
> Priority: Minor
> Labels: findbugs
> Fix For: 0.99.0, 0.98.5, 2.0.0
>
> Attachments: HBASE-11537-0.98.patch, HBASE-11537.patch,
> HBASE-11537.patch.txt
>
>
> In {{ProcedureCoordinator}} and {{ProcedureMember}} we synchronize on an
> instance of {{ConcurrentMap}} instead of using the interface methods for
> dealing with concurrent access.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
