[
https://issues.apache.org/jira/browse/HDDS-7061?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18022169#comment-18022169
]
Priyesh K commented on HDDS-7061:
---------------------------------
When nodes {{xxx-1.com}} and {{xxx-9.com}} went down, the system triggered a
*Pipeline CLOSE* command to SCM. This happens according to the configured
{{{}hdds.command.status.report.interval{}}}, which means there is always a
delay before the change is reflected.
For example, when you run:
{{ozone admin container info <container_id>}}
the command retrieves pipeline information from SCM. Until SCM receives the
*Pipeline CLOSE* and updates the status, it continues to show the old state.
This delay is expected, as the command does not check directly at the DN side
but instead fetches data from SCM. The SCM itself only updates at the
configured reporting interval, which causes the lag in reflecting the actual
state.
So closing this ticket as Resolved.
> ozone admin container info container_ID displays wrong replica info while EC
> offline reconstruction is still going on
> ---------------------------------------------------------------------------------------------------------------------
>
> Key: HDDS-7061
> URL: https://issues.apache.org/jira/browse/HDDS-7061
> Project: Apache Ozone
> Issue Type: Bug
> Reporter: George Huang
> Assignee: Priyesh K
> Priority: Major
>
> Repro:
> # ozone admin container info container_ID
> # kill some EC replica
> # rerun command: ozone admin container info container_ID. It will display
> old replica info.
> # It may take a while before this command displays correct/new replica info.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
