[ https://issues.apache.org/jira/browse/HDFS-8999?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14729407#comment-14729407 ]
Daryn Sharp commented on HDFS-8999: ----------------------------------- My rhetorical question to provoke thought was misunderstood. I was not implying we should eliminate IBRs. They do fix data inconsistency and ensure the NN stays in sync with the DNs state. I would never rely on the client as an authoritative source of what's on a DN. bq. So one of the motivations for this simplification is to decrease the total number of messages sent from DNs to NN. Recently we saw in a real cluster because a large amount of writing was happening, the IBR from DNs finally overwhelmed NN and caused congestion of the RPC queue. Happens all the time. Ironically I've been testing a patch to aggregate IBRs. No pipeline or DN changes necessary. > Namenode need not wait for {{blockReceived}} for the last block before > completing a file. > ----------------------------------------------------------------------------------------- > > Key: HDFS-8999 > URL: https://issues.apache.org/jira/browse/HDFS-8999 > Project: Hadoop HDFS > Issue Type: Bug > Components: namenode > Reporter: Jitendra Nath Pandey > > This comes out of a discussion in HDFS-8763. Pasting [~jingzhao]'s comment > from the jira: > {quote} > ...whether we need to let NameNode wait for all the block_received msgs to > announce the replica is safe. Looking into the code, now we have > # NameNode knows the DataNodes involved when initially setting up the > writing pipeline > # If any DataNode fails during the writing, client bumps the GS and > finally reports all the DataNodes included in the new pipeline to NameNode > through the updatePipeline RPC. > # When the client received the ack for the last packet of the block (and > before the client tries to close the file on NameNode), the replica has been > finalized in all the DataNodes. > Then in this case, when NameNode receives the close request from the client, > the NameNode already knows the latest replicas for the block. Currently the > checkReplication call only counts in all the replicas that NN has already > received the block_received msg, but based on the above #2 and #3, it may be > safe to also count in all the replicas in the > BlockUnderConstructionFeature#replicas? > {quote} -- This message was sent by Atlassian JIRA (v6.3.4#6332)