[infinispan-issues] [JBoss JIRA] (ISPN-2729) [FineGrainedAtomicMap] Class Cast Exception in Clustered + Repeatable Read + Write Skew

Galder Zamarreño (JIRA) jira-events at lists.jboss.org
Mon Jan 28 12:38:47 EST 2013


    [ https://issues.jboss.org/browse/ISPN-2729?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12750839#comment-12750839 ] 

Galder Zamarreño commented on ISPN-2729:
----------------------------------------

Pedro, where are these tests coming from? Do you have a particular use case for this?

There's currently no support for versioned delta-aware cache entries. Given that concurrent version updates could be possible within fine grained maps, it'd need some careful thought and testing.

For 5.2, we could add a message/error to indicate the unsupported nature of this combination of settings.
                
> [FineGrainedAtomicMap] Class Cast Exception in Clustered + Repeatable Read + Write Skew
> ---------------------------------------------------------------------------------------
>
>                 Key: ISPN-2729
>                 URL: https://issues.jboss.org/browse/ISPN-2729
>             Project: Infinispan
>          Issue Type: Bug
>          Components: Core API, Fine-grained API, Locking and Concurrency
>         Environment: Mac OSx 10.5, Java 1.6 sun-jdk, maven 3.0.4
>            Reporter: Pedro Ruivo
>            Assignee: Galder Zamarreño
>              Labels: atomic_map, clustered, write-skew
>             Fix For: 5.2.0.Final
>
>
> Test case available in here [1]
> When a transaction commits, the write skew check performs the validation by casting the cache entry to a ClusteredRepeatableReadEntry. However, the entry is a DeltaAwareCacheEntry.
> This happens when you have operations over a map inside a transaction.
> The tests failing are: 
> WriteSkewFineGrainedAtomicMapAPITest.testSizeOnCache()
> WriteSkewFineGrainedAtomicMapAPITest.testCreateMapInTx()
> [1]https://github.com/pruivo/infinispan/commits/t_atomic_map_issue

--
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