[ https://issues.apache.org/jira/browse/HDDS-642?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16701147#comment-16701147 ]
Ajay Kumar commented on HDDS-642: --------------------------------- [~linyiqun] thanks for updating the patch. Patch looks very good to me. Few NITs: # SCMChillModeManager ** L168, shall we mention in javadoc that this function is only for testing purpose. ** L79 either we should either avoid instantiating PipelineChillModeRule or fail if pipelineManager is null. ** Since in a real cluster we may not have a open pipeline in beginning, this rule should be turned off by default.How about a config to turn it on/off? (Default value to be off, we can enable it explicitly for smoke tests) # TestScmChillMode ** L254 By default chill mode is turned on so this should be true. Shall we replace this wait logic with assert? > Add chill mode exit condition for pipeline availability > ------------------------------------------------------- > > Key: HDDS-642 > URL: https://issues.apache.org/jira/browse/HDDS-642 > Project: Hadoop Distributed Data Store > Issue Type: Bug > Components: SCM > Reporter: Arpit Agarwal > Assignee: Yiqun Lin > Priority: Major > Attachments: HDDS-642.001.patch, HDDS-642.002.patch, > HDDS-642.003.patch, HDDS-642.004.patch, HDDS-642.005.patch > > > SCM should not exit chill-mode until at least 1 write pipeline is available. > Else smoke tests are unreliable. > This is not an issue for real clusters. -- This message was sent by Atlassian JIRA (v7.6.3#76005) --------------------------------------------------------------------- To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org