[ https://issues.apache.org/jira/browse/HDFS-4389?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13551602#comment-13551602 ]
Todd Lipcon commented on HDFS-4389: ----------------------------------- Hey Daryn. I vaguely remember this being a conscious decision at some point, but maybe I made that up. Two thoughts that might be relevant: 1) TestPersistBlocks seems to have started failing much more often recently, but I don't have evidence for this. Any chance something else might have caused a regression here? 2) In the old code, which retried over the restart, wouldn't it end up just hitting a SafeModeException and then failing at that point, when the NN was restarted? Given that the NN usually takes 30+seconds to leave safemode after starting, any retrying clients would probably hit that and fail anyway, no? > Non-HA DFSClients do not attempt reconnects > ------------------------------------------- > > Key: HDFS-4389 > URL: https://issues.apache.org/jira/browse/HDFS-4389 > Project: Hadoop HDFS > Issue Type: Bug > Components: ha, hdfs-client > Affects Versions: 2.0.0-alpha, 3.0.0 > Reporter: Daryn Sharp > Priority: Critical > > The HA retry policy implementation appears to have broken non-HA > {{DFSClient}} connect retries. The ipc > {{Client.Connection#handleConnectionFailure}} used to perform 45 connection > attempts, but now it consults a retry policy. For non-HA proxies, the policy > does not handle {{ConnectException}}. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira