[
https://issues.apache.org/jira/browse/CASSANDRA-10708?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15013780#comment-15013780
]
Yuki Morishita commented on CASSANDRA-10708:
--------------------------------------------
Thanks for the patch.
One thing that needs to be improved is that we can get {{null}} when doing
{{cfs.directories.find(filename)}}, and eventually causes NPE.
Can you update the branch with above fixed?
> Add forceUserDefinedCleanup to allow more flexible cleanup for operators
> ------------------------------------------------------------------------
>
> Key: CASSANDRA-10708
> URL: https://issues.apache.org/jira/browse/CASSANDRA-10708
> Project: Cassandra
> Issue Type: Improvement
> Components: Tools
> Reporter: Jeff Jirsa
> Assignee: Jeff Jirsa
> Priority: Minor
>
> {{nodetool cleanup}} currently executes in parallel on all sstables in a
> table. No source sstables are GCd until the parallel operation completes. In
> certain scenarios, this is nonideal (it has both memory and disk usage
> implications for operators who try to run the operation on larger tables).
> Adding {{forceUserDefinedCleanup}} puts cleanup operations closer to parity
> with compaction {{forceUserDefinedCompaction}} for the rare cases where
> operators need to do something slightly different than the traditional
> cleanup.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
