[
https://issues.apache.org/jira/browse/HDDS-1291?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Ajay Kumar updated HDDS-1291:
-----------------------------
Priority: Blocker (was: Major)
> Set OmKeyArgs#refreshPipeline flag properly when client reads a stale pipeline
> ------------------------------------------------------------------------------
>
> Key: HDDS-1291
> URL: https://issues.apache.org/jira/browse/HDDS-1291
> Project: Hadoop Distributed Data Store
> Issue Type: Bug
> Reporter: Xiaoyu Yao
> Assignee: Xiaoyu Yao
> Priority: Blocker
>
> After HDDS-1138, the OM client will not talk to SCM directly to fetch the
> pipeline info. Instead the pipeline info is returned as part of the
> keyLocation cached by OM.
>
> In case SCM pipeline is changed such as closed, the client may get invalid
> pipeline exception. In this case, the client need to getKeyLocation with
> OmKeyArgs#refreshPipeline = true to force OM update its pipeline cache for
> this key.
>
> An optimization could be queue a background task to update all the
> keyLocations that is affected when OM does a refreshPipeline. (This part can
> be done in 0.5)
> {code:java}
> oldpipeline->newpipeline{code}
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
