HP ML350 iLO: Cache Module Status Degraded

[D0natas]

Hello,

I'm installing and configuring two HP ML350e gen8 servers with aditional P420 RAID cards. I have configured two RAID1 logical drives. I have installed Citrix Xenserver software and get Cache Module Status Degraded message in the iLO.

How to figure out what is really causing this alarm?

I have enabled caching on logical drives, disabled write cache when capacitor not present and disabled physical drive write cache. 

[Casper42]

Don't know if SSA (replacement for ACU) supports Citrix Xen or not.

Reboot to Intelligent Provisioing and there is a full version of ACU or SSA (depends on Int Prov version) and you should be able to see the root cause there. Perhaps your cache module is just bad.

[Sbrown]

leave physical drive write cache on, or you will feel huge pain.

I've seen this when you put the P420 in the wrong slot, and overheat it.

I've seen this when you don't firmware up (cache board match P420).

Someone sold me a P420 blue-board with pre-release firmware on it. With a "RIG WIRE" - the same error occurred always , it just reduced the speed.

P420's fail was heat. It is why it works with some creativity with G6/G7 servers, but heat heat heat.

[D0natas]

I couldn't find SSA on Citrix.

When I boot to ACU, the cache module shows no errors. Everything is ok.

But when I boot to Xenserver again, after some time Cache Module Status shows degradation.

[D0natas]

It looks like the problem is because of overheat. The PCI slot shows 85 degrees Celsius. Changing slot doesn't help. The problem exists and slot temperature is the same.

I'm trying to increase cooling in the server bios, but the problem is that server is very loud.

[Casper42]

What firmware on the P420?

[D0natas]

I have tried 3.42 and 4.68.

The increase cooling didn't help. The cache degraded alarm appeared again. Only the PCI slot temperature decreased from 85 degrees Celsius to 53 degrees Celsius.

[D0natas]

The problem solved.

Capacitors were bad. I got new super caps from the support and now everything works ok. It's strange that two new servers had this problem.