HP T600 reboot for unknow reason!!

[Amnat Chuch]

Greetings,

I'm currently encountering the problem
with our HP 9000 T600 model. Since it
restart the machine itself for unknown
reason. My engineer checked in the SP
log and found only the msg below

" PSCM detects a +/- 16V or +/- 12V
failure CRITICAL
MPM status given below
final status: + 5 BS MASTER; +2B MASTER
-2.5B MASTER; +2C MASTER, -2.5C MASTER;
+3T MASTER; under current;"

And it did happened twice in one week.
Could anyone tell me
1. what does the message mean, and is it likely or unlikely?
2. what should we do to prevent its auto restart?
3. Does any patch required to fix the problem (we're currently using hp-ux 10.20?
4. Any hardware error and how to check on?

Your assistance will be deeply appreciated.

Regards,

Amnat

[Eugeny Brychkov]

Amnat,
looks like power (supply) failure. Do you have server backed up by UPS? Can there be anything wrong on the input AC line?
How system behaves when restarting? Does it restart immediately after failure? Do you have to powercycle it?
Eugeny

[Amnat Chuch]

Eugeny,

Thanks for your answer. But we already checked the power supply, and there was no problem. Also we do have UPS backed up for our machine. Anyway, I will check with the input AC line if there would be anything wrong.

The way the reboot happends is when the PSCM found that error, then the Service Processor (SP)do its self-test and restart system immediately. Our big problem is we do not know when this would happended, and it causes some damages after reboot. Would the problem be with the SP itself?

By the way, what do you mean by powercycle it?

Thanks.

Amnat

[Eugeny Brychkov]

At the console press Ctrl-B combination to get to Service processor. You'll see SP> prompt. Capture the following outputs: DL (display logs), EN (environment), FE (AC status), FV (firmware revisions), PS (power status) and attach them to your next reply
Eugeny

[Eugeny Brychkov]

Ahh... forgot... after outputs gathering to return to console mode type CO
Eugeny

[Patrick Wallek]

Have you checked /etc/shutdownlog to see if anything was logged there?

Do you have a hardware support contract on the machine? If so, call HP and have them look at it.

[T G Manikandan]

You should also check your /var/adm/syslog/OLDsyslog.log for any errors.


REvert

[Amnat Chuch]

Hi,

Thanks for all replied messages. My engineer notices that the operating status of the Service processor keeps going through the sequence of 00 11 7F repeatedly. Does anyone knows the meaning of these codes? It should be found in the CE handbook, but unfortunately we don't have this book and the service contract with HP.

Regards,

Amnat

[Eugeny Brychkov]

Attach outputs I requested in my previous reply. Attach what your engineer found too
Eugeny

[Patrick Wessel]

Amnat,
The first message you posted (PSCM detects...) means, that the Power System Control Monitor (PSCM) detects a problem with the Miscellaneous Power Module (MPM). The PSCM is located on the Service Processor card and shows it???s status on the two 7 segment displays on the back of the SP.
That fits perfectly to the error codes in your last response, because 00 is always the first code issued but 11 means that the PSCM shut down the power system (usually because of a problem)
You deal definitive with a power problem! It can be either with the MPM or with a device powered by the MPM. It would be helpful to know the status of the LEDs on the MPM (it???s installed in the front).

[Amnat Chuch]

Patrick,

Thanks for your information. We did changed the MPM, but the SP opearating status remains unchanged. So, we suspect in the I/O device powered by MPM such as the SCSI card. Is there any way to check it up before replacing it? And by the way, did you know about the "7F" SP code I mentioned last time? This might be a clue for the error device.

Thanks,

Amnat

[Patrick Wessel]

Happy New Year Amnat,
That is bad news about your MPM. Did the ???old one??? indicate any error status or were all LEDs green?
Usually it doesn???t make any sense to decode the so-called trail codes from the PSCM because the codes are already decoded in the SP logs. I do not know the meaning of 7F. But, if I???m not completely wrong does the MPM not power any interface cards. It should be responsible to power the SP (and PSCM). So, which came first: the hen or the egg? It might be a malfunctioning PSCM or the lines between SP and MPM are bad, which leads to two suspects: the SP or the backplane. Replacing the backplane doesn???t sound very attractive to me (and I haven???t seen many defect backplanes in T-Class systems). Do you have an other SP in spare?

[erics_1]

Amnat,

Does you system generate any ts99 files in /var/tombstones? If so, post the file so we can take a look.

Thanks,
Eric

[Amnat Chuch]

Hi

Sorry for no response, we???d checked with the ground circuit of the computer room and all over to ensure that there is no electrical power problem, and we found no errors.
Finally, we planned to change the SP on this weekend. Unfortunately, we are currently using SP with firmware ver. 2.43, but we could find the spare part with firmware ver. 2.40 Will it be any significant problem in terms of system operation if we replace the part with older firmware version? Or how can we upgrade our firmware version? Where? Is it free of charge, or charge? Any information will be deeply appreciated. We are so in rush. Thanks.

Amnat