[
https://issues.apache.org/jira/browse/HDFS-5124?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13747694#comment-13747694
]
Chris Nauroth commented on HDFS-5124:
-------------------------------------
bq. Do we even need a read lock on the namespace at all? Token verification has
nothing to do with the namespace.
The only potential problem I see with removing the namesystem read lock is that
failover acquires the write lock in {{HAState#setStateInternal}}. If this NN
is in the middle of transitioning from standby to active, then holding the read
lock in {{DelegationTokenSecretManager#retrievePassword}} blocks clients until
the transition to active has completed. Without the read lock, clients will
get an immediate {{StandbyException}} even though this NN is about to become
active. If this happens rapidly enough, then the client could exhaust its max
retries and get an RPC error before the transition to active completes.
I kicked off a Jenkins run for the new patch:
https://builds.apache.org/job/PreCommit-HDFS-Build/4870/
> 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: Jing Zhao
> 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
