[
https://issues.apache.org/jira/browse/HDFS-1172?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13009403#comment-13009403
]
Hairong Kuang commented on HDFS-1172:
-------------------------------------
I worked on a similar solution for our internal branch. Let me explain what I
did. Assume that a block's replication factor is r. When a block under
construction is changed to be complete, if it has r1 finalized replicas and r2
unfinalized replicas, NN puts r2 replicas into pending queue. If r1+r2<r, NN
also puts the block into the neededreplication queue. Does this algorithm make
sense?
> Blocks in newly completed files are considered under-replicated too quickly
> ---------------------------------------------------------------------------
>
> Key: HDFS-1172
> URL: https://issues.apache.org/jira/browse/HDFS-1172
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: name-node
> Affects Versions: 0.21.0
> Reporter: Todd Lipcon
> Attachments: HDFS-1172.patch
>
>
> I've seen this for a long time, and imagine it's a known issue, but couldn't
> find an existing JIRA. It often happens that we see the NN schedule
> replication on the last block of files very quickly after they're completed,
> before the other DNs in the pipeline have a chance to report the new block.
> This results in a lot of extra replication work on the cluster, as we
> replicate the block and then end up with multiple excess replicas which are
> very quickly deleted.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
