HPE EVA Storage

Connectivity problem between Alpha ES40 and EVA4000

 
SOLVED
Go to solution
Valentin Likoum
Frequent Advisor

Connectivity problem between Alpha ES40 and EVA4000

Hello all, we have a strange situation here:

Two Alphas ES40 (prod and standby), two EVA4000 (both prod). XCS version 5.031. Normally EVAs presents disks to the prod system only. Before switching prod to standby Alpha, I present disks from EVA to the standby (going-to-be-prod) system and unpresent them from the used-to-be-prod system after the swithing completed. I did it many times and it always worked as expected. A week ago we changed some failed parts on the standby system (FCAs remains the same) and I upgraded the firmware (SRM) on the standby ES40 from 7.0-2 to 7.3-1 (the main suspect but I can't prove it). Firmware on FCAs (DS-A5132-AA aka FCA-2684) was also changed from 1.90A4 to 1.91X6. Now the upgraded standby system can't see the EVA's disks neither from console mode nor from VMS.

"wwidmgr -show wwid" returns nothing.

"wwidmgr -show port -full" returns info about ports of FCAs and ports of FC swithes FCAs are connected to only.

No errors reported.

What have been done:

- unpresent disks on both EVAs from stanby system and present them again

- remove standby system from Host list and re-add it back. EVAs does see the standby system' ports and offers them in drop-down menu while adding host. So I concluded that standby system' ports are working  normally and they are visible from EVAs

- reboot one of the FC swithes (2/16V B-Series) the standby system is connected to

- reboot one of the EVAs

- downgrade firmware on ES40 from 7.3-1 back to 7.0-2 (FCA' firware was also downgraded)

- play with FFAUTO and FFNEXT console vars

All of above have no success and I have no more ideas to try. Do you have any?

1 REPLY 1
Valentin Likoum
Frequent Advisor
Solution

Re: Connectivity problem between Alpha ES40 and EVA4000

Forget it, guys. The reason was as dumb as possible. I exchanged everything except the most obvious and easiest to exchange one - the optical cables between two FCAs. And they were the reason.

It's still hard to understand why deleting and re-adding host did not have an effect but now it's just a rhetorical question.