vCenter Upgrade with RP4VMs

Keith, we will be publishing vsphere and VC upgrade procedures soon. Please send me details around the version and procedure you followed.

Thanks, Z

Thanks Z,

I will send you the details. I opened a case and the result was to re-install the cluster for fastest resolution so I am rebuilding it now. I'll shoot you an email soon depending on the re-install with the case number and all that.

Thanks!

-Keith

Hi Zahid,

Along the same lines of a vcenter upgrade - is it known/expected that an ESX host upgrade on a target site host would trigger a full init? I saw this a while back with a VIB upgrade going to P1 and I just got asked by a customer that saw this same phenomenon when they upgraded their ESX hosts and I'm betting it is related, but not 100% sure if it is expected or not. When I did the P1 VIB upgrade and saw it, I had a case open already which is what prompted the P1 upgrade as the fix and the response then was that it wasn't expected...but I could trigger the issue at will a couple times in that scenario. I'll test this out going to SP1 in the coming days, but figured I would ask ahead of time since it might be a few days before I can do it myself.

Thanks!

-Keith

Hi Keith,

When upgrade the vSphere version on ESXi, the node needs to be removed from the cluster, upgraded and then added back into the cluster.

Thanks,

Zahid

Twitter: @zDataGuy

Hi Keith,

Did you invoke maintenance mode to move the VMs to another ESXi host prior to the upgrade?

Regards,

Rich

Hi Guys,

I will hit them both here in this reply:

Rich - Yup, that was done to clear off the host and do their ESX upgrade

Zahid - I will check with the customer, but I am not 100% sure if the host was removed from the cluster prior to upgrade, I think it was just cleared off and put into maintenance mode. If it is not removed, would the upgrade trigger the init given the VIB is affected upong upgrade/reboot and relates to the ESX cluster as the splitter?

Thanks!

-Keith

Zahid - after talking to the customer this morning, it appears they did follow the process you describe above - they evacuate the VM's from the host, put in maintenance mode, move to their "upgrade cluster" and then perform the ESX upgrade. So it appears that even in doing that they still triggered these full init's which doesn't sound expected - have you seen any of that in your testing?

-Keith

Keith, what do you mean by "move to upgrade cluster"?

Here is a run down of the process being qualified - plz note this is tentative and should only be followed once published (there are also other approaches)

a. enter ESX to maintenance mode

b. remove ESX from ESX cluster

c. upgrade the ESX with splitter

d. return ESX to ESX cluster

e. get ESX out of maintenance mode

If they followed the above and still incurred a full sweep, we need to check the RPVM and vsphere versions, please have them open an SR and send me the # offline.

Thanks, Z

Sent from my iPhone

Yup...so that is what they did. When I say "move" I was talking about them taking it out of their "prod" cluster and putting it into a new cluster they create just called "upgrade" - more for management purposes, but it meets the steps b-d above according to what they told me when i inquired.

I'll inquire about the SR...they might be all done now and just took the hit so not sure what they will want to progress with, but I will ask given this was done a few days ago. Thanks!

Update - looks like this is a known issue/process as this was just posted today: https://support.emc.com/kb/476282

-Keith