hadoop-common-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "dhruba borthakur (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HADOOP-3503) Race condition when client and namenode start block recovery simultaneously
Date Thu, 05 Jun 2008 23:14:47 GMT

    [ https://issues.apache.org/jira/browse/HADOOP-3503?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12602843#action_12602843
] 

dhruba borthakur commented on HADOOP-3503:
------------------------------------------

This LOG calls occur only when the lease is recovered. This shud not happen too frequently.
I would like to have some log messages to debug this code when it it occurs. Do you see a
LOG message that occurs outside of lease recovery?

> Race condition when client and namenode start block recovery simultaneously
> ---------------------------------------------------------------------------
>
>                 Key: HADOOP-3503
>                 URL: https://issues.apache.org/jira/browse/HADOOP-3503
>             Project: Hadoop Core
>          Issue Type: Bug
>          Components: dfs
>    Affects Versions: 0.18.0
>            Reporter: dhruba borthakur
>            Assignee: dhruba borthakur
>         Attachments: pingPong.patch
>
>
> When a client detects a error while writing to a block, it starts the generation-stamp-protocol
to remove stale replicas. At the same time, if the namenode experiences a lease expiry event
for that file, the namenode starts the generation-stamp-protocol for the same block. Now,
the client and thr namenode ping-pongs trying to stamp the block replicas. This ping-pong
can continue for a long time.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message