[infinispan-issues] [JBoss JIRA] (ISPN-2435) Cache#replace(key, old, new) method doesn't check old value on non-local invocations
Sanne Grinovero (JIRA)
jira-events at lists.jboss.org
Tue Jan 22 14:45:21 EST 2013
[ https://issues.jboss.org/browse/ISPN-2435?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12749557#comment-12749557 ]
Sanne Grinovero commented on ISPN-2435:
---------------------------------------
[~anistor] I have an old test that I had given to Mircea, but it was failing because of a list of issues, not least it was the one highlighting the need for locking in order only the primary node.
Have a look into my github branch "ReplaceOperationStressTest"; it would fail for multiple reasons. I just rebased it but couldn't run it, no clue how it behaves today but I'm confident it will still fail, or Mircea would have resolved this ;)
I don't remember the full list but it was failing because of at least these bugs:
- it would return true even on failing compare operations
- if the key is not owned locally it would replace the value without checking for equality of _old_
- it would often deadlock (I believe this is solved with the new locking pattern implemented by Mircea after our design in Palma)
> Cache#replace(key, old, new) method doesn't check old value on non-local invocations
> ------------------------------------------------------------------------------------
>
> Key: ISPN-2435
> URL: https://issues.jboss.org/browse/ISPN-2435
> Project: Infinispan
> Issue Type: Bug
> Components: Core API, Distributed Cache
> Affects Versions: 5.1.8.Final, 5.2.0.Beta2
> Reporter: Sanne Grinovero
> Assignee: Mircea Markus
> Priority: Critical
> Fix For: 5.3.0.Final
>
>
--
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
More information about the infinispan-issues
mailing list