[
https://issues.apache.org/jira/browse/HDFS-15169?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17073378#comment-17073378
]
Ayush Saxena commented on HDFS-15169:
-------------------------------------
bq. Do we need to replace it at Router when receive result from NN? I am not
sure if it is graceful.
I think there is no other option, other than replacing at router, Agreed, that
might not look very graceful giving away the NN path won't be correct.
bq. About failover, is it proper to fix by retry?
Check the logic in {{RouterRpcClient}}, if the same can be pulled here, if not
we can ensure the basic functionality atleast that the other namenodes in the
namespace are tried.
One more issue I feel in this, Just for knowledge, is the FSCK output correct
incase of nested mount points, say mnt/dir is a mount point and FSCK on /mnt,
if /mnt is mounted and when /mnt is not there. If this isn't very correct, we
can handle this too as a follow up, as this also won't be too trivial to handle
here
Ideally we should have marked RouterFSCK as EVOLVING or something since it
isn't completed in full, to have the liberty of incompatible changes. but seems
to late now, branch-3.3 has already been created.
> RBF: Router FSCK should consider the mount table
> ------------------------------------------------
>
> Key: HDFS-15169
> URL: https://issues.apache.org/jira/browse/HDFS-15169
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Components: rbf
> Reporter: Akira Ajisaka
> Assignee: Xiaoqiao He
> Priority: Major
> Attachments: HDFS-15169.001.patch, HDFS-15169.002.patch,
> HDFS-15169.003.patch, HDFS-15169.004.patch, HDFS-15169.005.patch
>
>
> HDFS-13989 implemented FSCK to DFSRouter, however, it just redirects the
> requests to all the active downstream NameNodes for now. The DFSRouter should
> consider the mount table when redirecting the requests.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
