[
https://issues.apache.org/jira/browse/HDFS-729?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12769902#action_12769902
]
dhruba borthakur commented on HDFS-729:
---------------------------------------
> Is this a regular fsck with less output? That might still be prohibitively
> long and expensive for regular poll
Yes, at this point, I am visualizing it as a regular fsck with less output.
The problem with making this a new Namenode RPC is that this RPC would have an
upper limit on the number of corrupted files that can be returned via one
single invocation of the RPC. This kind-of- reduces the elegance of such an
API. The alternative is to make this new RPC retrieve a max number of corrupted
files together with a cookie that can be used in the next invocation of the RPC
to retrieve the remaining set of corrupted files (similar to readdir).
If we use a regular fsck, it does not lock the NN for an extended period of
time, neither does it have a problem if the number of files to be retrieved is
huge.
> fsck option to list only corrupted files
> ----------------------------------------
>
> Key: HDFS-729
> URL: https://issues.apache.org/jira/browse/HDFS-729
> Project: Hadoop HDFS
> Issue Type: Improvement
> Reporter: dhruba borthakur
> Assignee: dhruba borthakur
>
> An option to fsck to list only corrupted files will be very helpful for
> frequent monitoring.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
