HPE 3PAR 7450 - IOs stuck in ptbl on all the 4 nodes

[abhilash2]

HPE 3PAR 7450, 4 nodes. OS version: 3.2.1.200(MU2);P07,P10,P11,P12

We have a situation where IOs are stuck in ptbl on all of the 4 nodes. Does anyone else encounter this situation? I contacted HPE 3PAR support, once they said patch 80 fixes it. Then another support person said it was an oversight by the first support engineer. The fix is only for 3.2.1 MU3 and not for 3.2.1 MU2. The only fix is a downtime with reboot of all 4 nodes.

We caught the issue during OS upgrade - pre-upgrade health check:

Started Processing check_stuck_ptbl

Node0 WARNING: 4 IOs stuck in ptbl. Reboot node0 before upgrade.

Node1 WARNING: 3 IOs stuck in ptbl. Reboot node1 before upgrade.

Node2 WARNING: 2 IOs stuck in ptbl. Reboot node2 before upgrade.

Node3 WARNING: 5 IOs stuck in ptbl. Reboot node3 before upgrade.

We CANNOT afford to take a downtime! How can a storage vendor expect a customer to take downtime? Talk about 4 nodes and true HA!!

Anyone else in a similar situation? Any fix?

[Dennis Handly]

Can you reboot the nodes one at a time?

Also, Support should re-check to make sure these aren't false positives.

[Cali]

3.2.1 MU2 is from 1/2014 and Inactive since 7/2015.

Most Cars see more Service than Central SAN Storages.

Have you try to Reboot Node by Node?

[Dennis Handly]

> 3.2.1 MU2 is from 1/2014

I think the issue is that they can't upgrade due to this error.

[abhilash2]

Support clearly said we cannot reboot one-by-one because all the nodes have stuck IO. The only way out is bring down everything gracefully.

[abhilash2]

Exactly. We have upgraded the rest of 3PARs, just this last one is remaining with stuck IOs.

[abhilash2]

Wish we could do that, node-by-node reboot option is ruled out.