[
https://issues.apache.org/jira/browse/HDFS-14045?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16677179#comment-16677179
]
Erik Krogen commented on HDFS-14045:
------------------------------------
Hey folks, I think the current patch is an improvement, but I have one
modification I'd like to suggest / discuss. Currently the patch separates out
the metrics into active / non-active. However, in trunk we allow multiple
standbys, and after HDFS-12943 we will allow active, standby, and observer. It
would be useful to have a separate metric to monitor each. For example, on
trunk now, if the standby metric is unusually high, does this mean all of the
standbys are performing poorly? If not, how do I figure out which one? It would
be helpful to add the NameService / NameNode ID into the name of the metric,
then an operator can easily see the performance of each.
> Use different metrics in DataNode to better measure latency of
> heartbeat/blockReports/incrementalBlockReports of Active/Standby NN
> ----------------------------------------------------------------------------------------------------------------------------------
>
> Key: HDFS-14045
> URL: https://issues.apache.org/jira/browse/HDFS-14045
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: datanode
> Reporter: Jiandan Yang
> Assignee: Jiandan Yang
> Priority: Major
> Attachments: HDFS-14045.001.patch, HDFS-14045.002.patch,
> HDFS-14045.003.patch
>
>
> Currently DataNode uses same metrics to measure rpc latency of NameNode, but
> Active and Standby usually have different performance at the same time,
> especially in large cluster. For example, rpc latency of Standby is very long
> when Standby is catching up editlog. We may misunderstand the state of HDFS.
> Using different metrics for Active and standby can help us obtain more
> precise metric data.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
