[
https://issues.apache.org/jira/browse/HDDS-731?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16662712#comment-16662712
]
Xiaoyu Yao commented on HDDS-731:
---------------------------------
Thanks [~msingh] for the patch. It looks good to me overall. Here are a few
comments:
HddsDatanodeService.java
Line 95: NIT: The HDDSDatanodeShutdownHook class can be simply into a simple
lambda ()->stop()
Shall we put the priority to higher like 30, which is typically used for
similar services such as HDFS Datanode Block Pool, etc.
OzoneHddsDatanodeService.java
Line 76: we should set handler = null after line 76.
VolumeSet.java
Since we are removing the volume usage save shutdown hook here. I assume this
will be taken care by the chain of the ServicePlugIn shutdown. Can you clarify
the flow of service shutdown?
Note when the whole JVM is shutdown, we only worry about critical update
missing such as volume usage, buffer flush, or ratis snapshot as mentioned in
the description. Not all routinely shutdown sequences have to be executed.
Otherwise, it is very likely to timeout without all the operations being
executed.
> Add shutdown hook to shutdown XceiverServerRatis on daemon stop
> ---------------------------------------------------------------
>
> Key: HDDS-731
> URL: https://issues.apache.org/jira/browse/HDDS-731
> Project: Hadoop Distributed Data Store
> Issue Type: Bug
> Components: Ozone Datanode
> Reporter: Mukul Kumar Singh
> Assignee: Mukul Kumar Singh
> Priority: Major
> Attachments: HDDS-731.001.patch
>
>
> Currently on shutting down a Ozone datanode using "ozone --daemon stop
> datanode", the XceiverServerRatis is not shutdown properly. This jira
> proposes to add a shutdownHook to take ratis snapshot on shutdown
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
