[
https://issues.apache.org/jira/browse/HDFS-5124?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13747808#comment-13747808
]
Daryn Sharp commented on HDFS-5124:
-----------------------------------
bq. Shouldn't a properly tuned listen backlog prevent OOM though? Or have you
seen one of those cases where the OS doesn't really enforce the listen backlog
you requested?
The OS backlog will just artificially throttle the RPC server a bit. The
listener thread is draining the backlog as fast as it can, and the OS is in
turn filling it. The server does try to close some connections on each loop,
but on each loop it's sucking the listen queue dry and creating {{Connection}}
objects for every accepted connection and assigning them to socket readers
blocked on the FSN lock - which will prevent even admin commands from getting
in.
> Namenode in secure cluster deadlocks
> ------------------------------------
>
> Key: HDFS-5124
> URL: https://issues.apache.org/jira/browse/HDFS-5124
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: namenode
> Affects Versions: 2.1.1-beta
> Environment: Secure Hadoop 2 cluster
> Reporter: Deepesh Khandelwal
> Assignee: Daryn Sharp
> Priority: Blocker
> Attachments: HADOOP-5124.patch, HDFS-5124.001.patch,
> HDFS-5124.002.patch, nn_jstack.out
>
>
> Namenode deadlocks after a while in use.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
