[
https://issues.apache.org/jira/browse/GEODE-3519?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16146128#comment-16146128
]
ASF subversion and git services commented on GEODE-3519:
--------------------------------------------------------
Commit 2637bd8794b76e048f2fb13007ad0a50842059d8 in geode's branch
refs/heads/develop from [~bschuchardt]
[ https://git-wip-us.apache.org/repos/asf?p=geode.git;h=2637bd8 ]
GEODE-3519 servers are not locking on some ops initiated by clients
While investigating dlocktoken cleanup I discovered that a number of
operations coming from a client were not locking entries for Scope.GLOBAL
regions on servers. Only put, putIfAbsent and variants of replace were
obtaining locks.
> servers are not locking on remove or invalidate ops initiated by clients
> ------------------------------------------------------------------------
>
> Key: GEODE-3519
> URL: https://issues.apache.org/jira/browse/GEODE-3519
> Project: Geode
> Issue Type: Bug
> Components: client/server
> Reporter: Bruce Schuchardt
> Assignee: Bruce Schuchardt
>
> While debugging a DLockToken leak I found that a Scope.GLOBAL region in a
> server was recording CachePerfStats.conflatedEvents. This should be zero in
> a Scope.GLOBAL region because each operation on the region is supposed to be
> performed while holding a distributed lock on the affected key.
> The problem stems from DistributedRegion not overriding the "basicBridge"
> operations that handle these events and surrounding the superclass's method
> with locking.
> The only operations currently performing correct locking are "put" and
> "replace".
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
