[
https://issues.apache.org/jira/browse/CASSANDRA-5379?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13611860#comment-13611860
]
Jonathan Ellis commented on CASSANDRA-5379:
-------------------------------------------
We TTL hints with gcgs to prevent this:
{code}
. // The hint TTL is set at the smallest GCGraceSeconds for any of the
CFs in the RM;
// this ensures that deletes aren't "undone" by delivery of an old hint
int ttl = Integer.MAX_VALUE;
for (ColumnFamily cf : mutation.getColumnFamilies())
ttl = Math.min(ttl, cf.metadata().getGcGraceSeconds());
{code}
> Do not deliver hints older than a configured age
> ------------------------------------------------
>
> Key: CASSANDRA-5379
> URL: https://issues.apache.org/jira/browse/CASSANDRA-5379
> Project: Cassandra
> Issue Type: Improvement
> Components: Core
> Affects Versions: 1.2.3
> Reporter: amorton
> Priority: Minor
>
> From http://www.mail-archive.com/[email protected]/msg28682.html
> There exists a possibility for a hinted insert created before a delete to be
> delivered after a tombstone that would have covered the insert is purged. I
> don't think it's very likely but it's possible.
> Could we filter the RowMutation to remove CF's where the gc_grace is less
> than the age of the hint? Or add a config property for a max_hint_age?
--
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
