MSA2012i single controller

[Guy Newman]

Hi

We have an urgent problem with a single controller iSCSI san. This has been working fine with no errors, it has now been shipped to Germany and put live, and since has been losing connectivity to its controller every few hours when under load. After a physical reboot it is possible to reconnect the iSCSI and see all the drives. In the event log i see the following critical events prior to failure

07-02 11:30:07 107 A231 Critical Error: Fault Type: NMI p1: D000000 p2:20000000 p3: 0200003 p4:0000000 CThr: Host

C 07-02 11:30:07 107 A230 Critical Error: Fault Type: NMI p1: 02E90FF p2:01F5808 p3: 01631EB p4:01529D9 CThr: Host

Can anyone shed any light on what this means?

[Johnny von Heimerall]

Hi Guy,

Verify that you are running the latest firmwares (ie. J210P12 for the storage controller)
If you are, i suggest you open a case with HP for deeper analyzis.
NMI = Non maskable interrupt
(non-recoverable hardware error(s))
Most likely its a fw issue but could as well be a hardware.

Btw, are you running snaps on this machine?

[digone52]

Hi Guy.

We have had the same fault. Can you please inform us if this was a hardware or firmware problem?

Thanks.

[BSavage]

107
A317 Critical Error: Fault Type: NMI p1: 02E90FF p2:01F5808 p3: 01631EB p4:01529D9 CThr: Host


I just had the exact same error yesterday on my MSA 2012i. I have J210P12 firmware running. No other events showed any issues before this Critical error. The web management reported management controller could not communicate with the storage controller. Anyone have the answer to what the cause is?

Brian

[Guy Newman]

Hi all

This was confirmed as a firmware issue by HP, basically the firmware shipped with the controller had a known bug causing this problem.

Unfortunately we had to wait a few weeks for the patch to come out (as at that time they only had a patch for the sas/fibre controller versions) but there is now one available for iscsi. We are now running controller firmware version J210P19.

We have been soak testing this for a while and gradually adding more load. So far we have had no further issues.