Arun Sarin created HDDS-9782:
--------------------------------
Summary: Changing the Ozone service Id makes the cluster[OM] state
irrecoverable
Key: HDDS-9782
URL: https://issues.apache.org/jira/browse/HDDS-9782
Project: Apache Ozone
Issue Type: Bug
Reporter: Arun Sarin
Changing the Ozone service Id makes the cluster state irrecoverable
Scenario:
1. Change the Ozone Service ID from ozone1 to say ozone2
2. Restart the cluster.
3. The cluster fails to come up and throws an error.
Now will the cluster go in unrecoverable state?
Change the service id back to original [Ozone 1] that will be allowed or its
not allowed ?
Repro Steps to be Followed :
1. Change the Ozone Service ID , say from ozone1 to say Ozone2.
2. Restart the cluster.
3. The cluster fails to come up and throws an error.
4. Revert the Ozone Service ID name to what it was [Ozone1].
5. Do a cluster restart.
6. OM doesn't start back again & cluster goes in bad state.
Note : We update only ozone.om.service.id .
ozone.metadata.dir and ozone.om.db.dir were not updated.
Also it was observed that
{code:java}
There are 2 ratis group directories created here which should not happen
[root@host-1 hadoop-ozone]# cd /var/lib/hadoop-ozone/om/ratis
[root@host-1 ratis]# ls
a870c7c2-351a-3867-aca7-b9a5b4809b76 d39ebeec-41e8-35f1-a92b-9f198c4c3682
{code}
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
