[
https://issues.apache.org/jira/browse/CASSANDRA-5335?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13600015#comment-13600015
]
Dean Hiller commented on CASSANDRA-5335:
----------------------------------------
that was supposed to read "brought it to it's knees" not "needs"
> need ability to switch to LCS node by node
> ------------------------------------------
>
> Key: CASSANDRA-5335
> URL: https://issues.apache.org/jira/browse/CASSANDRA-5335
> Project: Cassandra
> Issue Type: Bug
> Affects Versions: 1.2.2
> Environment: linux, cassandra 1.2.2
> Reporter: Dean Hiller
> Labels: compaction
>
> when we switched to LCS, nearly all the servers started rewriting their file
> structure. We need a way to switch just one node so we can slowly roll out a
> huge change like this. doing the whole cluster brought it to it's needs and
> we tried nodetool stop compaction on some nodes but that did not help at all.
> We had to really wait overnight for all LCS to finish.
> We had also tried nodetool setcompactionthroughput 1 which also had no effect
> and performance was still pretty bad. Ideally, none of this would be a
> problem if we were just doing one node as we would just pull the node back
> out while it was converting to LCS.
--
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
