[ https://issues.apache.org/jira/browse/CASSANDRA-3550?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13160957#comment-13160957 ]
Sylvain Lebresne commented on CASSANDRA-3550: --------------------------------------------- What about a way to force marking down a node through JMX? Usually, node gets down during the distributed tests only when we kill them. There is no point in that case to wait even a few seconds for other to notice, and on the other side I'm afraid that having a phi threshold too low could result in test flickering in some case (but maybe I worry too much). Of course, if we add a JMX call, we want to make sure we have a few tests that actually test the 'real life' failure detector behavior. > Relax phi threshold enforcement > ------------------------------- > > Key: CASSANDRA-3550 > URL: https://issues.apache.org/jira/browse/CASSANDRA-3550 > Project: Cassandra > Issue Type: Improvement > Components: Core > Reporter: Brandon Williams > Assignee: Brandon Williams > Priority: Trivial > Fix For: 1.0.6 > > > Currently we clamp the phi threshold between 5 and 16. For distributed > testing purposes, 5 slows things down. Instead let's allow values down to 1, > but add a warning for anything less than 5. -- 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