StoreVirtual 12 Failed Upgrade

[AuZZZie]

Trying to upgrade a 2 node P4300 G2 to OS 12. Both nodes return 100% diagnostic tests and everything checks out but during the upgradde I receive the following. 

Gathering the installation definitation information for SA-D-sAN1-N2 failed. The upgrade partition is currently in use. Canceling all further installations. 

Tried rebooting the nodes. No difference. 

Anyone seen this?

[Davo666]

I'm in a similar situation. Have a Management Group with a pair of 4300 G2's, a pair of 4330's and a pair of 4350's. All devices were on 10.5. Now have the pair of 4300's on 12.0, one of the 4350 on 12.0 but the other stuck on 11.0, and both the 4330's stuck with "Upgrade partition in use" like you. I found a post elsewhere which indicated a potential fix which I am doing now. Reboot one of the nodes, allow volumes to re-sync, then shut-down and resart the CMC. Rinse and repeat for the other node (or all remaining nodes in any cluster, ensuring the CMC is restrtaed each time).

Will let you know how I go.

[Davo666]

Nope ! Fail

Just keeps going in circles throwing up the name of rnadom nodes in the cluster each time.

Guess I'm calling HP support tomorrow morning. Saw another post where Support was able to clear / delete the upgrade partition.

[AuZZZie]

Keep us posted how you get on. We don't have active support on ours as we basically gave up on it with the stagnant software updates and long delays. 

[s_mcleod]

It looks like mtab gets corrupted during the upgrade and thus the sans fail to unmount the upgrade partition.

We've given up on HP support as we've found them pretty useless and they've caused more trouble than they're worth.

If you look around the internet you can find a way to get root on the SANs, it's not at all hard and takes less time than HP 'technicians' trying and getting the password wrong from their generator app.

ps -ef | grep -i upgrade

kill the upgrade process thats broken

umount /mnt/lhnupgrade

reboot

[shabba_1]

Same issue trying to go from 11.5 to 12.5. Reboot of all nodes did not sort issue. Not sure how to get root console on the nodes. Seems that killing a process and doing a unmount should be done by a reboot though..

[Marelven]

Hello Davo666,

Same problem here with 2 P4300 nodes.

Did you have any luck with HP support? If not, did you find something else?

Regards

[Marelven]

In this doc, http://h20565.www2.hpe.com/hpsc/doc/public/display?sp4ts.oid=5306917&docId=emr_na-c03983796&docLocale=en_US, there is a troubleshooting section where this problem is described.

To resolve this, we're supposed to do the following but I'm not sure I get it:

It is important to perform the following steps on all of the storage systems that might be affected; otherwise the next attempt will fail, putting storage systems that have just been corrected back into a bad state.

On each storage system affected, close the CMC and then re-open it. This is necessary; otherwise, the next attempt will fail, putting storage systems that have just been corrected back into a bad state

Shouldn't a reboot of each node do exactly this, close the CMC? Is there something I'm missing?


Regards