[
https://issues.apache.org/jira/browse/HDFS-15127?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17018519#comment-17018519
]
Ayush Saxena commented on HDFS-15127:
-------------------------------------
Thanx [~elgoiri] for the update, Seems the test failure is due to changes in
{{checkDirectoriesFaultTolerant}}, as we didn't change any logic in the
directory creation part, revoking the changes in
{{checkDirectoriesFaultTolerant}} would make the test work.
I don't think we can do similar stuff in directory creation too, as for
directory creation, {{mkdirs}} will be called as part of {{invokeConcurrent}},
so even if one subcluster was down, the directory would have been created in
other subcluster. So there would be no point failing the operation, when
directory has already been created.
> RBF: Do not allow writes when a subcluster is unavailable for HASH_ALL mount
> points.
> ------------------------------------------------------------------------------------
>
> Key: HDFS-15127
> URL: https://issues.apache.org/jira/browse/HDFS-15127
> Project: Hadoop HDFS
> Issue Type: Improvement
> Reporter: Íñigo Goiri
> Assignee: Íñigo Goiri
> Priority: Major
> Attachments: HDFS-15127.000.patch, HDFS-15127.001.patch,
> HDFS-15127.002.patch
>
>
> A HASH_ALL mount point should not allow creating new files if one subcluster
> is down.
> If the file already existed in the past, this could lead to inconsistencies.
> We should return an unavailable exception.
> {{TestRouterFaultTolerant#testWriteWithFailedSubcluster()}} needs to be
> changed.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
