[
https://issues.apache.org/jira/browse/HADOOP-2346?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12574034#action_12574034
]
Raghu Angadi commented on HADOOP-2346:
--------------------------------------
> The class SocketIOWithTimeout has two different uses, reading and writing.
> Why not create two classes?
That's why we have SocketInputStream and SocketOutputStream, right? I don't
know advantage of two more classes on top of 3 we already have.
SocketIOWithTimeout is just a helper class for reading and writing with a
timeout. Its is not user facing. I still don't follow the problem completely.
It will be best if you could modify the patch to how you wanted it to be.. that
way I understand what you mean.
> DataNode should have timeout on socket writes.
> ----------------------------------------------
>
> Key: HADOOP-2346
> URL: https://issues.apache.org/jira/browse/HADOOP-2346
> Project: Hadoop Core
> Issue Type: Bug
> Components: dfs
> Affects Versions: 0.15.1
> Reporter: Raghu Angadi
> Assignee: Raghu Angadi
> Fix For: 0.17.0
>
> Attachments: HADOOP-2346.patch, HADOOP-2346.patch, HADOOP-2346.patch,
> HADOOP-2346.patch, HADOOP-2346.patch, HADOOP-2346.patch, HADOOP-2346.patch,
> HADOOP-2346.patch, HADOOP-2346.patch, HADOOP-2346.patch, HADOOP-2346.patch
>
>
> If a client opens a file and stops reading in the middle, DataNode thread
> writing the data could be stuck forever. For DataNode sockets we set read
> timeout but not write timeout. I think we should add a write(data, timeout)
> method in IOUtils that assumes it the underlying FileChannel is non-blocking.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
