hyperscale x os migration from RHEL to Rocky experiences

That should say Hyperscale OS* - I’ll get it fixed :)

The upgrades have been going smoothly from where I sit. We have a considerations & walkthrough upgrade video that we should be posting next week.

I had issues with it partly because documentation wasn’t clear and partly because we had borrowed index cache to increase /ws/ddb sizes which messed with the extended attributes. I have our lab upgraded but haven’t tried in production yet. We also have a new cluster running Rocky with no issues so far.

@Chris Colhard what was missing from documentation?

I assumed you put in maintenance mode first like all other procedures but you do not. It said you could run in ssh using tmux but you cannot, you must use ILO/IPMI.

Hi all,

In the documentation, it stated that “Copy the latest HyperScale X image from Commvault store to the /ws/ddb/iso folder on the Hyperscale X remote software cache node.” Do you guys have the iso directory on the cache node? I didn’t have it, do I need to created iso directory? 

Yes create the iso directory then copy the iso over to it.

Now they are saying having the iso in /ws/ddb is wrong, they are saying to mount using the ilo then boot to the iso via the ilo until they have a fix which is supposed to be out next week (Feb 10, 25)

great timing as I am just about to start my upgrade.

Hi ​@Chris Colhard 

Could you point to the case where they mentioned? I have worked on multiple migrations and didn’t had to do so.

Regards,

Case 241205-803

Thanks ​@Chris Colhard. Yes, for HSX setup you have, manual mount needs to be done. We are seeing these in some cases and not needed for all HSX Cluster.

Regards,

Update, our issue is because of nvme drives on HPE Apollos...it just took me approx 28 hours to migrate 4 nodes from RHEL to Rocky using the iso=true method. Now i’m being told a fix will be in a new MP (11.32.95?) and a new ISO will be out but haven’t gotten dates yet.