[
https://issues.apache.org/jira/browse/HDFS-11021?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16868969#comment-16868969
]
Konstantin Shvachko commented on HDFS-11021:
--------------------------------------------
I think separating IBR and FBR lock metrics out of OTHER will be valuable for
monitoring cluster health. This are internal (not client-facing) operations, so
increase in latency can be treated as an alert that something is going wrong on
the cluster.
> Add FSNamesystemLock metrics for BlockManager operations
> --------------------------------------------------------
>
> Key: HDFS-11021
> URL: https://issues.apache.org/jira/browse/HDFS-11021
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Components: namenode
> Reporter: Erik Krogen
> Assignee: Erik Krogen
> Priority: Major
>
> Right now the operations which the {{BlockManager}} issues to the
> {{Namesystem}} will not emit metrics about which operation caused the
> {{FSNamesystemLock}} to be held; they are all grouped under "OTHER". We
> should fix this since the {{BlockManager}} creates many acquisitions of both
> the read and write locks.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
