HPE SimpliVity

Dell OmniCube CN-3000 Accelerator Card change?

Go to solution

Dell OmniCube CN-3000 Accelerator Card change?

Hi. I

I swapped the faulty accelerator card with one from a working node. It seems that it actually could work.

All lights are green and $SVTLOG complains about wrong NVRAM UUID.

I’m no exprert but as much as I undertand, this NVRAM is basically the block map or something like that?

If I guess correctly, I should just delete the NVRAM on the TIA card or load the NVRAM from the disk to the TIA to overwrite the current one.

Any ideas anyone?


Output from $SVTLOG:

2020-09-10T22:47:34.505Z INFO  0x7f80050e3840 [:] [nvrammanager] nvram_manager.cpp:1516 Disk  Label  Consistent: 1

2020-09-10T22:47:34.505Z INFO  0x7f80050e3840 [:] [nvrammanager] nvram_manager.cpp:1517 Disk  Sequence   Number: 70

2020-09-10T22:47:34.505Z INFO  0x7f80050e3840 [:] [nvrammanager] nvram_manager.cpp:1518 Disk  Unique Identifier: c2a50d2b-6dbc-49b3-a134-9226036e2dca

2020-09-10T22:47:34.505Z INFO  0x7f80050e3840 [:] [nvrammanager] nvram_manager.cpp:1519 NVRAM Label  Consistent: 1

2020-09-10T22:47:34.505Z INFO  0x7f80050e3840 [:] [nvrammanager] nvram_manager.cpp:1520 NVRAM Sequence   Number: 12

2020-09-10T22:47:34.505Z INFO  0x7f80050e3840 [:] [nvrammanager] nvram_manager.cpp:1521 NVRAM Unique Identifier: 7f14b0fa-8074-4f6a-8f64-cd2a316fac20

2020-09-10T22:47:34.505Z FATAL 0x7f80050e3840 [:] [nvrammanager] nvram_manager.cpp:1533 Disk and NVRAM UUID mismatch -- cannot start NVRAM Manager.

2020-09-10T22:47:34.505Z FATAL 0x7f80050e3840 [:] [nvrammanager] nvram_manager.cpp:2779 NVRAM manager failed start.

2020-09-10T22:47:34.505Z WARN  0x7f80050e3840 [:] [nvrammanager] nvram_manager.cpp:2838 Skipped write to disk due to no Disk Write Object.

Valued Contributor

Re: Dell OmniCube CN-3000 Accelerator Card change?

The replacement of the Accelerator card in a SimpliVity system requires updates to the stored identities and drive signatures to allow the system to start with access to the existing data.

That will require a support engagement to complete those actions.

Alternatively, the system can be removed from the federation and redeployed.

If there are other systems in the same cluster as the failed system, the data will be rebuilt automatically after the redeployment. If no other systems exist in the same cluster, data would need to be restored from backups sent to other clusters.

Otherwise a redeployment will result in data loss.

I am an HPE employee.