[
https://issues.apache.org/jira/browse/HDDS-14063?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Siddhant Sangwan resolved HDDS-14063.
-------------------------------------
Resolution: Not A Bug
> Under replicated Quasi_Closed containers are created on write at disk space
> boundary
> ------------------------------------------------------------------------------------
>
> Key: HDDS-14063
> URL: https://issues.apache.org/jira/browse/HDDS-14063
> Project: Apache Ozone
> Issue Type: Sub-task
> Components: Ozone Datanode
> Affects Versions: 2.1.0
> Reporter: Siddhant Sangwan
> Assignee: Siddhant Sangwan
> Priority: Major
>
> Suppose there's a 6 Datanode cluster, where 4 datanodes have reached almost
> full disk capacity so that there's no space for even one more container. 2
> datanodes are close to full and have disk space for some more blocks.
> In this case, clients can still write new Ratis containers such that it's
> created on the two datanodes but not on the third datanode in the pipeline
> (because of no remaining space). This container will be QUASI_CLOSED and will
> be permanently under replicated until more space is made available.
> This is because Ozone allows majority commit. This is not a bug, just an
> observation made during testing that I'm documenting.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
