[
https://issues.apache.org/jira/browse/HBASE-952?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jim Kellerman resolved HBASE-952.
---------------------------------
Resolution: Fixed
Fixed. Committed.
> Deadlock in HRegion.batchUpdate
> -------------------------------
>
> Key: HBASE-952
> URL: https://issues.apache.org/jira/browse/HBASE-952
> Project: Hadoop HBase
> Issue Type: Sub-task
> Affects Versions: 0.19.0
> Reporter: Jim Kellerman
> Assignee: Jim Kellerman
> Priority: Blocker
> Fix For: 0.19.0
>
>
> {code}
> public void batchUpdate(BatchUpdate[] bus, Integer[] locks)
> {code}
> Takes out splitsAndClosesLock before calling checkResources. If
> checkResources stalls waiting for a cache flush, and a split starts before
> the flush, HRegion.close deadlocks when it tries to acquire
> splitsAndClosesLock
> This never resolves because batchUpdate is waiting for a flush that will
> never happen since close has already disabled them.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
