[
https://issues.apache.org/jira/browse/HDDS-5136?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Arpit Agarwal updated HDDS-5136:
--------------------------------
Issue Type: Improvement (was: Bug)
> Recycling of pipelines in ozone
> -------------------------------
>
> Key: HDDS-5136
> URL: https://issues.apache.org/jira/browse/HDDS-5136
> Project: Apache Ozone
> Issue Type: Improvement
> Components: SCM
> Reporter: Shashikant Banerjee
> Assignee: Shashikant Banerjee
> Priority: Major
>
> Currently, in ozone pipelines can remain open for ever, unless there is a
> failure detected. The idea here is to close pipelines over a times and
> recycle them to achieve the below objectives:
> 1) With pipelines, being active raft log disks can get full and eventually
> fail well before closing. As the pipelines can not be closed, the raft log
> disk cannot be cleaned up. Autoclosing over a time, will help keep the raft
> log disk space maintainable.
> 2) Pipelines with datanodes with slow disk/bad disk/full disk can be closed
> actively and then moved out of pipeline allocation.
> 3) In cases such node addition, disk additions , pipelines on nodes with
> lower disk capacity can be closed and new IO's can be redirected to new
> nodes/disks.
> 4) Will help the cluster make more rack and toplogy aware over time.
> With multiraft, all the above challenges/challenges become more significant.
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
