Cache battery?

There is no cache battery on this raid card, this is super capacitors. 

Do you have an error in WRKPRB ? If you have 2 cards you may be have a failing card, or failing capacitor, data are not cached in memory and are read write directly to the disk. Like When card have cache battery.

Ding ding ding! Virgile wins!  It's super capacitors!

From the ticket:

REPORTING SERIAL: YL10UE645049 (Based on the serial number this is a topaz card 2016 / May)

-

There is a brief window of time when the Super Capacitors are charging and if the Super Capacitor status is checked during the brief window of time, the problem occurs.

For adapters manufactured before 4th quarter 2019, the adapter needs to be replaced although the adapter firmware needs to be installed through the PTFs listed. If a new adapter with the Diamond capacitor is installed in the managed system, these PTFs ensure the new adapter also gets the new microcode update.

The adapter firmware installs with a system IPL or with an IOP reset of the adapter.

Adapter firmware level 19512C00 is provided by the following Hyper IBM i PTFs:

MF69409 for release 720

MF69402 for release 730

MF69408 for release 740

MF69410 for release 750

Let this be a warning to all those who are thinking of not renewing maintenance on their Power 9 after the warranty runs out.  We keep maintenance.  I'd like to replace a different sandbox P10 with that new model and replace this P9 with that other P10.

Dear Robert

I figure you may have known by now what the problem is.  But if not, your WRKDSKSTS output that you posted looks like a case of the failure of a disk unit in a RAID-5 (or 6) set.  Based on my limited experience with this disk HW problem, if the problem is in your disk controller card, some disk units in WRKDSKSTS should NOT still be active but some of them still are active in your case.  Are all shown as degraded by now? If so, then the problem can be in the card such as write cache failure, in which case the cards' microcode disables caches in both cards and continue working without the caches.

BTW, do you use HDD or SSD?  If the latter, your case would be my first case of SSD unit failure for IBM i internal disk!  

------------------------------
Satid S
------------------------------

Satid, I've lost numerous SSD's.  I was less than impressed with them when the first models came out.  You paid more for them.  You paid extra for maintenance for them (as they were not included on system maintenance like spinning disks were).  And, they didn't get the performance I expected.  I got much better bang for the buck on a system with EXTENSIVE journalling by purchasing the HA journal performance option.  I think that option is free now.