[ https://issues.apache.org/jira/browse/HDFS-2981?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13213103#comment-13213103 ]
Todd Lipcon commented on HDFS-2981: ----------------------------------- I disagree that it should be true by default. Apps like HBase which are latency-sensitive don't want to wait for a whole block to be re-transferred when a node in the pipeline fails. Apps which are long-running writes and non-latency-sensitive (eg log collection) can flip this to true on their own client, no? > The default value of > dfs.client.block.write.replace-datanode-on-failure.enable should be true > --------------------------------------------------------------------------------------------- > > Key: HDFS-2981 > URL: https://issues.apache.org/jira/browse/HDFS-2981 > Project: Hadoop HDFS > Issue Type: Improvement > Reporter: Tsz Wo (Nicholas), SZE > Assignee: Tsz Wo (Nicholas), SZE > Attachments: h2981_20120221.patch > > > There was a typo earlier in the default value of > dfs.client.block.write.replace-datanode-on-failure.enable. Then, HDFS-2944 > changed from "ture" to "false". It should be changed to "true". -- 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