[
https://issues.apache.org/jira/browse/HADOOP-2757?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12704218#action_12704218
]
Hairong Kuang commented on HADOOP-2757:
---------------------------------------
I do not understand the problem that this patch is trying to solve. You claimed
that this patch is to provide soft mount feature to dfs. But the major change
is made to LeaseChecker.
1. LeaseChecker does not get to run if there is no writer. So this patch does
not solve the problem if a no-writing call to dfs hangs.
2. In the writing case, if NN hangs, the call renew will hang too. I do not see
how this patch can detect the hanging NN.
That's why I suggest that the change should make to IPC Client when it sends
ping to detect if NN is alive after it does not get the response back in time.
The disadvantage of my proposal is that it will also effect Map/Reduce.
> Should DFS outputstream's close wait forever?
> ---------------------------------------------
>
> Key: HADOOP-2757
> URL: https://issues.apache.org/jira/browse/HADOOP-2757
> Project: Hadoop Core
> Issue Type: Improvement
> Components: dfs
> Reporter: Raghu Angadi
> Assignee: dhruba borthakur
> Attachments: softMount1.patch, softMount1.patch, softMount2.patch
>
>
> Currently {{DFSOutputStream.close()}} waits for ever if Namenode keeps
> throwing {{NotYetReplicated}} exception, for whatever reason. Its pretty
> annoying for a user. Shoud the loop inside close have a timeout? If so how
> much? It could probably something like 10 minutes.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
