[ https://issues.apache.org/jira/browse/HBASE-17287?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15945696#comment-15945696 ]
Enis Soztutar commented on HBASE-17287: --------------------------------------- bq. In patch v5, before starting the mini cluster, I set config for master not to host meta region. Ok makes sense. I've checked the test again, seems good. The timeout of 60 secs is aggressive I think though. Let's bump that to 3 mins. In Jenkins things can run super slow causing flakiness. On the other issue, do we have a problem with the default setup (master hosting meta table and regionserver) that master abort is not causing the daemon to go down? If we are good there, +1 for the patch. > Master becomes a zombie if filesystem object closes > --------------------------------------------------- > > Key: HBASE-17287 > URL: https://issues.apache.org/jira/browse/HBASE-17287 > Project: HBase > Issue Type: Bug > Components: master > Reporter: Clay B. > Assignee: Ted Yu > Fix For: 1.4.0, 2.0 > > Attachments: 17287.branch-1.1.v4.txt, 17287.branch-1.v3.txt, > 17287.branch-1.v4.txt, 17287.master.v2.txt, 17287.master.v3.txt, > 17287.master.v4.txt, 17287.master.v5.txt, 17287.v2.txt > > > We have seen an issue whereby if the HDFS is unstable and the HBase master's > HDFS client is unable to stabilize before > {{dfs.client.failover.max.attempts}} then the master's filesystem object > closes. This seems to result in an HBase master which will continue to run > (process and znode exists) but no meaningful work can be done (e.g. assigning > meta).What we saw in our HBase master logs was:{code}2016-12-01 19:19:08,192 > ERROR org.apache.hadoop.hbase.master.handler.MetaServerShutdownHandler: > Caught M_META_SERVER_SHUTDOWN, count=1java.io.IOException: failed log > splitting for cluster-r5n12.bloomberg.com,60200,1480632863218, will retryat > org.apache.hadoop.hbase.master.handler.MetaServerShutdownHandler.process(MetaServerShutdownHandler.java:84)at > org.apache.hadoop.hbase.executor.EventHandler.run(EventHandler.java:129)at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)at > > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)at > java.lang.Thread.run(Thread.java:745)Caused by: java.io.IOException: > Filesystem closed{code} -- This message was sent by Atlassian JIRA (v6.3.15#6346)