[ https://issues.apache.org/jira/browse/HDFS-10876?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15507539#comment-15507539 ]
Wei-Chiu Chuang commented on HDFS-10876: ---------------------------------------- INFO may be fine. However, I am seeing an IOE in balancer that disrupts normal balancer operation. It only occurs after several hours of balancing. 2016-05-14 08:56:27,636 WARN [pool-2302-thread-7] balancer.Dispatcher (Dispatcher.java:dispatch(325)) - Failed to move blk_1550828018_1100035582559 with size=167 from 10.17.1.1:1004:DISK to 10.17.1.2:1004:DISK through 10.17.1.1:1004: Can't re-compute encryption key for nonce, since the required block key (keyID=-327392027) doesn't exist. Current key: -327392024 Balancer aborts after five iterations if no blocks are moved. If you make it DEBUG it's going to be really hard to tell what's going on. From a support perspective it's going to generated many support tickets. > Dispatcher#dispatch should log IOException stacktrace > ----------------------------------------------------- > > Key: HDFS-10876 > URL: https://issues.apache.org/jira/browse/HDFS-10876 > Project: Hadoop HDFS > Issue Type: Improvement > Components: balancer & mover > Affects Versions: 2.6.0 > Reporter: Wei-Chiu Chuang > Assignee: Manoj Govindassamy > Priority: Trivial > Labels: newbie > Attachments: HDFS-10876.001.patch > > > This error logging should be improved. A warning log should record the > stacktrace as well. > {code:title=Dispatcher#dispatch} > try { > ... > } catch (IOException e) { > LOG.warn("Failed to move " + this + ": " + e.getMessage()); > ... > } > {code} -- This message was sent by Atlassian JIRA (v6.3.4#6332) --------------------------------------------------------------------- To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org