[
https://issues.apache.org/jira/browse/HDDS-9633?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Devesh Kumar Singh resolved HDDS-9633.
--------------------------------------
Resolution: Duplicate
> Recon Container Report mis-match with RM Admin CLI container report
> -------------------------------------------------------------------
>
> Key: HDDS-9633
> URL: https://issues.apache.org/jira/browse/HDDS-9633
> Project: Apache Ozone
> Issue Type: Improvement
> Components: Ozone Recon
> Reporter: Devesh Kumar Singh
> Assignee: Devesh Kumar Singh
> Priority: Major
>
> * Difference between statistics (MISSING, MIS-REPLICATED, OVER-REPLICATED,
> UNDER-REPLICATED) of unhealthy containers provided by Recon and statistics
> provided by
> {code:java}
> ozone admin container report
> {code}
> command.
> _*All below degraded states of containers are marked as *UNHEALTHY* in
> Recon.*_
> {noformat}
> *MISSING* - Recon identifies any container as missing where it just checks if
> no replica of a container is
> reported in any of the DNs container report.
> *UNDER-REPLICATED *- Recon supports only Ratis containers marks the container
> as UNDER-REPLICATED,
> if replication factor doesn't match based on reports from all DNs and replica
> count is lesser than replication factor, where as RM container report
> includes both EC and Ratis containers and RM marks any container as
> UNDER-REPLICATED
> if it has all replicas present but not having sufficient number of HEALTHY
> replicas or if a container has some replicas missing.
> *OVER-REPLICATED* - Recon marks a container as OVER-REPLICATED if multiple
> DN reports more than the required number
> of replicas for a container. However RM container report marks a container
> as OVER-REPLICATED on 2 conditions - one, if there are all healthy
> replicas in excess (more than required nodes), another ,
> if there are in unhealthy replicas in excess along with required number of
> healthy replicas.
> MIS-REPLICATED - Depending on placement policy, both Recon and
> RM container report marks a container as MIS-REPLICATED i
> f container is not placed as per placement policy.
> {noformat}
> So all in all, there are differences in a way how Recon identifies the
> degraded state (UNHEALTHY) of containers and how RM container report
> identifies various degraded states.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
