[
https://issues.apache.org/jira/browse/HDFS-17292?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
huangzhaobo99 updated HDFS-17292:
---------------------------------
Description:
# Add a new metric of the number of times the
DatanodeManager#slowPeerCollectorDaemon thread collects SlowNodes, and display
it in a map structure.
# The same SlowNode may always appear in the prod env, so when
slowPeerCollectorDaemon is turned on, record the number of times it has been
collected by the slowPeerCollectorDaemon thread. If the collection frequency is
too high, SRE or DEV need to repair the machine.
was:
# Add a new metric of the number of times the
DatanodeManager#slowPeerCollectorDaemon thread collects SlowNodes, and display
it in a map structure.
# The same SlowNode may always appear in the prod env, so when
slowPeerCollectorDaemon is turned on, record the number of times it has been
collected by the slowPeerCollectorDaemon thread. If the collection frequency is
too high, SRE or DEV need to repair the machine.
# The following figure shows the number of write request
threads(DataNodeWriteXceiversCount) for SlowNodes collected by the
slowPeerCollectorDaemon thread at different time periods.
(If "DataNodeWriteXceiversCount" is 0, there is no write request, indicating a
SlowNode)
> Show the number of times the slowPeerCollectorDaemon thread has collected
> SlowNodes.
> ------------------------------------------------------------------------------------
>
> Key: HDFS-17292
> URL: https://issues.apache.org/jira/browse/HDFS-17292
> Project: Hadoop HDFS
> Issue Type: New Feature
> Reporter: huangzhaobo99
> Assignee: huangzhaobo99
> Priority: Major
> Labels: pull-request-available
>
> # Add a new metric of the number of times the
> DatanodeManager#slowPeerCollectorDaemon thread collects SlowNodes, and
> display it in a map structure.
> # The same SlowNode may always appear in the prod env, so when
> slowPeerCollectorDaemon is turned on, record the number of times it has been
> collected by the slowPeerCollectorDaemon thread. If the collection frequency
> is too high, SRE or DEV need to repair the machine.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
