[ https://issues.apache.org/jira/browse/CASSANDRA-3868?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13214983#comment-13214983 ]
Peter Schuller commented on CASSANDRA-3868: ------------------------------------------- Being "under capacity" is often not a binary and non-changing thing. You may be close to capacity; you may have a burst of traffic due to real traffic patterns, or you may have a burst of traffic because of the behavior of some other piece of software, or maybe you are temporarily having problems because someone is reading at QUORUM, a large compaction is causing a node to be backed up, and another neighbor goes down so that dynamic snitching doesn't save you, etc. Maybe you suffered a power outtage and lots of nodes are coming back with a cold page cache. Maybe.... you get the idea. :) Actually though, it does strike me that replicate_on_write would be more useful as a per-node JMX tweakable setting than a schema attribute. Having to push through a schema change is something you want to avoid anyway, and you'd definitely like to be able to tweak this on a per-node basis. Ok - I have changed my mind, looking at it from this perspective. I suggest that we keep the ability in place by a simple JMX adjustable flag global to a node (I can submit the patch for that if people agree), and remove the support for it in the schema as originally suggested. That should further avoid the chance of it being seen as a supported/primary feature somehow, and makes it more safe and easy to adjust this in near-real time. > Remove or nullify replicate_on_write option > ------------------------------------------- > > Key: CASSANDRA-3868 > URL: https://issues.apache.org/jira/browse/CASSANDRA-3868 > Project: Cassandra > Issue Type: Improvement > Components: Core > Affects Versions: 0.8.0 > Reporter: Brandon Williams > Fix For: 1.1.0 > > Attachments: 3868.txt > > > My understanding from Sylvain is that setting this option to false is rather > dangerous/stupid, and you should basically never do it. So 1.1 is a good > time to get rid of it, or make it a no-op. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira