HPE Storage Users Group

Hmm pity. Interesting though, i'll see if i can figure out some details on it.

_________________
The goal is to achieve the best results by following the clients wishes. If they want to have a house build upside down standing on its chimney, it's up to you to figure out how do it, while still making it usable.

Here is something we received from HP when upgrading our 7450. I actually did the upgrade to our 7200 and found the MGT port went offline, so when changing the port to be 1000FD it came back. Also I updated to P07 which then put the cages in a degraded state about FW level (needed to be at 321a). For some reason it didn't automatically upgrade the cages, hmmm? When I called support they said I just needed to execute "upgradecage -a". Then it upgraded them. Nice documenting this in the release notes HP.


Please review the following points and confirm accordingly:

1. Ensure the Host and Array recommendations that was shared earlier & also attached with this email are implemented. Failure to follow the recommendations (e.g. MS MPIO patch) may cause host disconnect/upgrade issues.
2. Let us know the details of the Contact person who would be available to approve controller node reboots and validate host status during the upgrade.
3. The preferred mode of communication, email or phone?
4. Please confirm if you have any Cisco Catalyst/NX series or HP Procurve switch in the environment. If yes, the below workaround needs to be followed before the upgrade

For Cisco Catalyst/NX series switches:
The Switch ports connected to management ports of 3PAR arrays may disable during
InformOS 3.2.1 MU2 upgrade if connected to Cisco Catalyst/NX series switches with link-flap detection
enabled. To prevent this, you may suggest the customers to disable the link flap detection feature on
ports attached to 3PAR management ports by executing the interface configuration command “no
errdisable detect cause link-flap” before firmware upgrade. This command needs to be
executed on the Cisco Catalyst/NX series Switches

For HP Procurve Switches:
If the Network Switch connected to the Admin ports on 3PAR is a HP Procurve, kindly set the interface speed on Procurve Switch to 1000 MBPS. This should be done to avoid the loss of communication between 3PAR Nodes and SP during the upgrade to 3.2.1.MU2.Not setting the speed to 1000 MBPS will result in SP-InServ communication issues during the OS upgrade



So going through the upgrade the first time resulted in loss of network connectivity between the SP and the Inserv. I had opened up a case with HP, but then remembered something about the network and this particular upgrade. I changed the network port that the Inserv connects to to 1000FD and network connectivity came back. After the upgrade completed successfully I applied P07. After the upgrade completed successfully again, the array stated the drive cages were not at the correct FW level. Finally HP calls me this morning and says to run “upgradecage –a”, viola, the drive cages started to upgrade to the correct FW of 321a, just like the 7450.

Shouldn’t all this be in the release notes?????

Things get worse, we did another pair of arrays the the RC ports went to half duplex and started flapping. Seems this bug also can impact RC ports as well. Any of our RC ports in a Cisic fabric extender dropped until we locked the Cisoc and RC port to 1000FD, great QA HP, not like Cisco is an obscure network vendor.

Ha... My Cisco MDS san switches management ports can't even auto negotiate properly with our Cisco Nexxus FEXes... we have to hard code one side or the other. I wouldn't be so quick to point a finger at HP on this one.

_________________
Richard Siemers
The views and opinions expressed are my own and do not necessarily reflect those of my employer.

3.2.1 MU2 P10 & P11 is on ftp.usa.hp.com available

and a 3.2.1eMU2.
It's unclear to me which patches are actually included in it though.

_________________
The goal is to achieve the best results by following the clients wishes. If they want to have a house build upside down standing on its chimney, it's up to you to figure out how do it, while still making it usable.

Hi,

We have recently upgraded 3PAR OS from 3.1.2 to 3.2.1 P7, P10, P11.
Upgrade to 3.1.2 MU2 resulted in lost connection to the 3PAR management ports (as the settings of 100Mb/Full/Non-Neg were automatically applied onto the 3PAR mgmt ports).
Having changed the switch ports' settings accordingly, the connectivity to 3PAR mgmt was regained.

After patching to P10/P11, the mgmt ports' settings on the 3PAR were changed again (automatically) to 100/Half/Auto-Neg and we noticed the errors on Eth interfaces.

Finally, after configuring Auto-Neg on the switch (which resulted in 1000/Full/Auto-Neg on the 3PAR), the earlier errors disappeared.