[
https://issues.apache.org/jira/browse/CASSANDRA-6646?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Marcus Eriksson updated CASSANDRA-6646:
---------------------------------------
Attachment: 0001-make-CorruptSSTableException-an-FSError-to-be-able-t.patch
[~kohlisankalp] do you have a stack trace? It looks to me like we catch all
CorruptBlockExceptions and either rethrow them as FSError or
CorruptSSTableException
Attached patch makes CorruptSSTableException extend FSError to make the
uncaught exception handler call handleFSError for them as well.
> Disk Failure Policy ignores CorruptBlockException
> --------------------------------------------------
>
> Key: CASSANDRA-6646
> URL: https://issues.apache.org/jira/browse/CASSANDRA-6646
> Project: Cassandra
> Issue Type: Bug
> Reporter: sankalp kohli
> Assignee: Marcus Eriksson
> Priority: Minor
> Fix For: 2.0.7, 2.1 beta2
>
> Attachments:
> 0001-make-CorruptSSTableException-an-FSError-to-be-able-t.patch
>
>
> If Cassandra is using compression and has a bad drive or stable, it will
> throw an CorruptBlockException.
> Disk Failure Policy only works if it is an FSError and does not work for
> IOExceptions like this.
> We need to better handle such exceptions as it causes nodes to not respond to
> the co-ordinator causing the client to timeout.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
