[ https://issues.apache.org/jira/browse/HDFS-9535?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15051981#comment-15051981 ]
Masatake Iwasaki commented on HDFS-9535: ---------------------------------------- bq. For the first block, it is possible that NN receives 0 block-received msg from DN but still commit the block when the client tries to get the next block. In that case we will add the block into under-replicated queue instead of the pending queue, and block recovery will happen in the cluster. {{BlockManager#processMisReplicatedBlock}} will not process incomplete block. I think the case is that block is completed but there is still pending block-received msg. {code} if (!block.isComplete()) { // Incomplete blocks are never considered mis-replicated -- // they'll be reached when they are completed or recovered. return MisReplicationResult.UNDER_CONSTRUCTION; } {code} > Fix TestReplication#testNoExtraReplicationWhenBlockReceivedIsLate > ----------------------------------------------------------------- > > Key: HDFS-9535 > URL: https://issues.apache.org/jira/browse/HDFS-9535 > Project: Hadoop HDFS > Issue Type: Bug > Affects Versions: 2.8.0 > Reporter: Jing Zhao > Assignee: Mingliang Liu > Priority: Minor > > TestReplication#testNoExtraReplicationWhenBlockReceivedIsLate failed in > several Jenkins run (e.g., > https://builds.apache.org/job/PreCommit-HDFS-Build/13818/testReport/). The > failure is on the last {{assertNoReplicationWasPerformed}} check. -- This message was sent by Atlassian JIRA (v6.3.4#6332)