OS upgrade to 3.2.1 strange issue

[mchartier]

Hi,

I did the upgrade from 3.1.3 MU1 P09 to 3.2.1 MU2 today on one of our systems and I ran into this :

First, I've upgraded the SPOCC to 4.3.0.GA-24. Then 3.2.1 MU2 for the OS (Online). All the health checks run well and then the node 0 was upgraded, rebooted and then back online. When the task start to upgrade the node 1, I've receive an error failed message (error code = 1) and then the system display that the node 1 was manually shutdown. After that, no more trace of the node 1 in the system. In the update windows, I've seen connection lost with the management. The IP address of the management was unavailable for about a minute.

Finally after a couple of minutes, the node 1 came back so I've returned to the SPOCC in the update section and I check for the last update status, the system then displayed that an update was trying to be performed and then the upgrade continues and finished succesfully.

After that, the upgradecage -a was very long, about half an hour (I have two cages with 20 pd each). But it finished without problems. upgradepd -a went fine and quick.

Any idea what could happen? I've succesfully applied patches 07, 10, 11 and 12 after.

No downtime at all during all theses steps...but really strange.

I have another system to update in a couple of weeks..so I'm curious to know what happen.

Any idea?

Regards,

[jgmke]

I did the upgrade to 3.2.1 and experienced similar problems. I was unable to get the admithw command to work until after I had rebooted both nodes one extra time.

Strange issue indeed. I haven't had any problems with any upgrades, except the one time.

[Davidkn]

Worth noting that a cageuprade does take and very long time, so thats normal, its around 15-20 mins per shelf.

ive not had any issues going to 3.2.1

[hdtvguy]

I honestly have been less than impressed with 3par upgrades. 3.1.3 and 3.2.1 MU2 were both major disruptions for us and the team doing the upgrades for 3.2.1 left steps undone and unverified. HP has struggled integrating and support 3par and after 3 years I am about done. We will need some more storage in the next few months and we are going to explore options. I have an army of account people form HP, but in the end it is 3par poor QA and poor support that hurts them every time. Not sure why after almost 4 years this is so hard for HP. I know the product team is not happy when I post this stuff and at this point I don't care any more, I have seen enough lately to no longer be the huge 3par fan I used to be. SSMC has lost its way and other things I can not disclose, still a decent product, just quickly losing favor and it advantage for us.

[cali]

This seems to be a local Problem.
On many Customer Sites in the last Year I have seen a impressive experience with clearly structured procedure from the HP CSC COE - Global Deployment Center.
They give a lot of Information before and after the Upgrades and do a lot of checks, that I never have done, if I do the Upgrades on myself.
And I never had an Issue that is not solved and never see any Downtime.
For me the 3PAR Deployment Team is on of the best I know.

Now Back to the Origin:
I have had the same Issue and this seems to be a Bug in the MU2 Update.
If it is not self resolved the Update Team can fix it.
The Bug is fixed with MU2 Patch11 or EMU2.
124556
Resolves an issue that does not allow to connect to the management interface of the array after
upgrading to HP 3PAR OS 3.2.1 MU2. Before attempting the upgrade, existing connection issues
,if any, must be resolved by contacting HP support.
So it is better to Update to EMU2 and not to MU2.

[mchartier]

OK. Then if that system is now at 3.2.1 MU2 Patch 12. Do you recommend anything else?

I'll check for the EMU2 for my other system when I will update it.

[hdtvguy]

This seems to be a local Problem.
On many Customer Sites in the last Year I have seen a impressive experience with clearly structured procedure from the HP CSC COE - Global Deployment Center.
They give a lot of Information before and after the Upgrades and do a lot of checks, that I never have done, if I do the Upgrades on myself.
And I never had an Issue that is not solved and never see any Downtime.
For me the 3PAR Deployment Team is on of the best I know.
.

Then I must have bad luck because both my major upgrades they dropped the ball. 3.1.3 they took 15+ hours to restore replication and half of that time was finding people that knew what they were doing. For 3.2.1 on some arrays they forgot to mount srdata volume when done, others they did not update firmware for drives and others failed they did not catch.

[AdamKnox]

It was confusing for us too, as they don't really say anything on the software download portal.

3PAR_OS_release_3.2.1_MU2_software_QR482-11028.iso = Base 3.2.1 + MU2

3PAR_OS_release_3.2.1_EMU2_software_QR482-11039.iso = Base 3.2.1 + MU2 + Patch 11

[mchartier]

Thanks AdamKnox,

And If I'm right, the patchs are not cummulatives? So Patch 11 didn't fix what's in 07 and 10?

[AdamKnox]

The patches aren't cumulative, so yes -- you need to load all of them (at least the ones that haven't already been superseded). For 3.2.1 MU2, that amounts to P07 + P10 + P11 + P12.