[
https://issues.apache.org/jira/browse/CASSANDRA-7288?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Marcus Eriksson updated CASSANDRA-7288:
---------------------------------------
Attachment: 0001-make-sure-we-don-t-try-to-compact-already-compacting.patch
We add all sstables to a user defined compaction without checking if the
sstables are actually compacting.
This means that if a compaction finishes after we add the descriptor to the
user defined compaction, but before those sstables are marked compacting, we
can end up compacting a file that is gone from the tracker and cause this assert
> Exception during compaction
> ---------------------------
>
> Key: CASSANDRA-7288
> URL: https://issues.apache.org/jira/browse/CASSANDRA-7288
> Project: Cassandra
> Issue Type: Bug
> Reporter: Julien Anguenot
> Assignee: Marcus Eriksson
> Attachments:
> 0001-make-sure-we-don-t-try-to-compact-already-compacting.patch,
> compaction_exception.txt
>
>
> Sometimes Cassandra nodes (in a multi datacenter deployment) are throwing
> errors during compaction. (see attached stack trace)
> Let me know what additional information I could provide.
> Thank you.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
