[ https://issues.apache.org/jira/browse/HDFS-14528?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16926881#comment-16926881 ]
Ayush Saxena commented on HDFS-14528: ------------------------------------- On a cursory view, the patch looks good. minor nit: correct the javadoc {code:java} + /** + * Test that manual failover is successful with Observernode. + */ {code} In the test you are not using observer, but standby... > Failover from Active to Standby Failed > ---------------------------------------- > > Key: HDFS-14528 > URL: https://issues.apache.org/jira/browse/HDFS-14528 > Project: Hadoop HDFS > Issue Type: Bug > Components: ha > Reporter: Ravuri Sushma sree > Assignee: Ravuri Sushma sree > Priority: Major > Attachments: HDFS-14528.003.patch, HDFS-14528.004.patch, > HDFS-14528.2.Patch, ZKFC_issue.patch > > > *In a cluster with more than one Standby namenode, manual failover throws > exception for some cases* > *When trying to exectue the failover command from active to standby* > *._/hdfs haadmin -failover nn1 nn2, below Exception is thrown_* > Operation failed: Call From X-X-X-X/X-X-X-X to Y-Y-Y-Y:nnnn failed on > connection exception: java.net.ConnectException: Connection refused > This is encountered in the following cases : > Scenario 1 : > Namenodes - NN1(Active) , NN2(Standby), NN3(Standby) > When trying to manually failover from NN1 to NN2 if NN3 is down, Exception is > thrown > Scenario 2 : > Namenodes - NN1(Active) , NN2(Standby), NN3(Standby) > ZKFC's - ZKFC1, ZKFC2, ZKFC3 > When trying to manually failover using NN1 to NN3 if NN3's ZKFC (ZKFC3) is > down, Exception is thrown -- This message was sent by Atlassian Jira (v8.3.2#803003) --------------------------------------------------------------------- To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org