[
https://issues.apache.org/jira/browse/HADOOP-7163?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13078453#comment-13078453
]
Dave Thompson commented on HADOOP-7163:
---------------------------------------
It looks like the client code cleans up and is capable of handling connection
failures, ready for retry. Would RetryProxy be suitable for the particular
scenario you're thinking of?
> "java.net.SocketTimeoutException: 60000 millis timeout" happens a lot
> ---------------------------------------------------------------------
>
> Key: HADOOP-7163
> URL: https://issues.apache.org/jira/browse/HADOOP-7163
> Project: Hadoop Common
> Issue Type: Bug
> Components: ipc
> Reporter: Owen O'Malley
> Assignee: Devaraj Das
> Fix For: 0.20.204.0
>
>
> We don't have retries for the case where the secure SASL connection is
> getting created from the tasks. There is retry
> for TCP connections, but once the TCP connection has been set up,
> communication at the RPC layer (and that includes
> SASL handshake) happens without retries. So for example, a client's "read"
> can timeout.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
