upgrade questions from 7.3 to 8.x regarding MPHA and snapmirror

[richard_unsworth]

I Am looking to upgrade data ontap in the near future from 7.3 to 8.x.

We have two FAS3140 filers, one at Location:A and the other at our D/R site, Location:B - both filers have  give away / take back setup within themselves for high availability.

Regarding MPHA:

After running through upgrade advisor it states that both locations will fail the MPHA test - I dont think this is true especially as we just added a couple of drawers to one location and performed the give away take back operations and everything worked just fine with no hiccups.

Is this somthing to worry about when performing the upgrade?

Regarding SnapMirror:

The documentation states that you must upgrade your SnapMirror destination before upgrading the source machine to ensure that SnapMirror keeps working.

We have snap mirrors going in both directions, Location:A to Location:B and Location:B to Location:A so this cannot be done.

In this situation, would you perform a give away on the filer at Location:B and a give away on the filer at Location:A and then upgrade the target destination followed by the source filer, and the perform a take back on both systems before performing the same steps on Filer B at both sites?

If this isnt the right way to do it, are we forced to quiesce the mirrors and break them, perform the upgrade and then re-quiesce and snap resume the mirrors?

any advice would be much appreciated.

Best Regards

Rich

[scottgelb]

You can quiesce, but don't need to break the mirrors.. I wouldn't want to leave a system in failover too long when not necessary so would recomment quiescing the mirrors then resuming after upgrades at both sides are done...

For MPHA takeover/giveback still works... single path works with clustering so it sounds like MPHA isn't configured on all shelves.  I would check sysconfig -a, storage show disk -p, fcadmin device_map or sasadmin expander_map and make sure all shelves are showing 2 paths to each controller... If you can get your NetApp or partner SE to stop by they can run Wiregauge... they can run on the console, over the network or even import your autosupport and find where the MPHA issue is at.  Or open a support case and they can find why the advisor shows mpha has an issue... it could be just one loop or stack causing the issue.

[richard_unsworth]

Hi Scott,

Thanks for you reply on the netapp forums, I was going to reply over there but for some reason I couldn’t see how to do that.

So basically are you saying that even though I have snap mirrors going in both directions, as long as I quiesce all of the mirrors on Filer:A at the disaster recovery center, make sure the latest firmware for the unit, drawers and drives is already loaded onto the filer, upgrade to 8.x of ontap from 7.3 on Filer:A at the D/R site and then perform the same upgrade at the corporate location, I should be ok? Make sure the firmware / drawer shelves and hard drive firmware is up to date etc. and after verifying that all of those updates were ok, resume the snap mirrors and do the same for Filer:B at both sides? (using the give away and take back I assume before the upgrade and after the upgrade of Filer:A at both locations) ?

Thank you very much in advance.

Rich

[scottgelb]

Correct… as long as ONTAP matches the same major release you can resume the mirrors.. independent of disk and shelf firmware, but dependent on ONTAP version. Target has to be the same or higher major release…so after one location is updated you can resume the mirrors but probably makes sense to update both sides then resume them all depending on your timing.