topic Re: CPU missing in HPE 9000 and HPE e3000 Servers

CPU missing

Jeffrey F. Goldsmith — Wed, 12 Nov 2003 15:13:01 GMT

This morning when I started Glance and checked on my CPU's I only see 3 of them. Does anyone know how I can check to see what happened to the fourth CPU?

It is an L2000 with HP-UX 11.0

Re: CPU missing

Selvaraj — Wed, 12 Nov 2003 15:38:53 GMT

Hi,

If you have GSP access, then go to GSP prompt
and type "SS". You may see all 4 CPUs or 3
may be active. YOu will see the status of 4th one.

If you would have been rebooted the box, then
look for "syslog" file in /var/adm/syslog and
grep the term "processor". You can also use
"top" command to see the processors status.

Other method, use "cstm" command and map for "cpu" devices and run info to generate infolog. Then view them. You will get the answer in minute.

Reply me for any comments.

-Selvaraj

Re: CPU missing

Jeffrey F. Goldsmith — Wed, 12 Nov 2003 15:44:29 GMT

I was checking the GSP when I received your e-mail. With the SS command I see all four processors but the fourth processor (3) in installed but not configured.

Whats up with that?

What do i need to do to get it configured again?

Re: CPU missing

Jeff Schussele — Wed, 12 Nov 2003 15:52:44 GMT

Hi Jeff,

What you need to do is go back to the GSP & look at the error logs. I think you'll find that during one of the last (if not last) boots you'll see that it errored out on a POST & was deconfigured. The other possibility was that is was manually deconfigured from the BCH menu by someone.
If it errored out log a HW call with the RC.

HTH,
Jeff

Re: CPU missing

Selvaraj — Wed, 12 Nov 2003 15:53:47 GMT

Hi Jeff,

If you are not familiar with "cstm" command, then just type the following command to get info log.

=======================================
echo "selclass qualifier cpu;info;wait;infolog"|cstm |more
=======================================

It is one line command.

Thanks
Selva

Re: CPU missing

Michael Tully — Wed, 12 Nov 2003 15:55:26 GMT

It could be that the 4th CPU is in fact dead .... Have a look at the difference in the two tombstone files. (/var/tombstones/ts99 and ts98)

Re: CPU missing

Michael Steele_2 — Wed, 12 Nov 2003 15:56:40 GMT

Please attach /etc/shutdownlog or note any panic's.

Also check for ICOD.

# swlist -l fileset | grep -i -icod

And check STM for your missing CPU's status. Should be displayed here.

Re: CPU missing

Selvaraj — Wed, 12 Nov 2003 16:03:24 GMT

Hello Jeff,

It looks like hardware problem. Why don't you try one complete reboot before logging for hardware call. Shutdown the machine, power off and then power on again. Stop at boot prompt and check configuration/ information/ service menu. Then boot with primary root disk. This may helps you.

Good luck.

-Selva

Re: CPU missing

Jeffrey F. Goldsmith — Wed, 12 Nov 2003 16:14:45 GMT

I went out and checked /var/tombstones/ts98 and /ts99.

ts98 had all 4 CPU's and ts99 only had 3 CPU's.

I think that I am going to have to report this as a bad processor.

Re: CPU missing

Jeffrey F. Goldsmith — Wed, 12 Nov 2003 16:39:32 GMT

i checked the /ect/shutdownlog but didnt find anything (no panics).

I have to contact all the departments that are using this box to schedule a reboot.

Anything else I can check for?

Re: CPU missing

Saurav_1 — Thu, 13 Nov 2003 07:48:02 GMT

Hi,

UR problem is resolved. The time stamp of ts99 file will tell when the proc was disconfigurred. contact HP for replacement. More over Pls do assign points accordingly.

Saurav

Re: CPU missing

Michael Steele_2 — Thu, 13 Nov 2003 09:13:07 GMT

I think Mr. Goldsmith already knows this since its in his posting.

Re: CPU missing

Stefan Stechemesser — Thu, 13 Nov 2003 10:36:56 GMT

Hi,

the Event Monitoring Services (EMS) deconfigures a CPU if there have been too many correctable errors (single bit cache errors) to avoid a possible crash due to a multibit error. The CPU is soft deconfigured online (no more processes will start on this CPU) and marked to be hard-deconfigured during the next reboot (=> the error happened before your last reboot). You should check the file /var/opt/resmon/log/event.log. There may be a message telling you why this CPU was deconfigured. It may also be (less probable) that the CPU failed during the selftest and deconfigured itself. Then you have nothing in the event.log file, but maybe in the GSP error log. You can reenable the CPU by rebooting, press any key within 10 seconds on the console after the selftests and then enter the Service Menu by entering "ser" at the BCH prompt. Then you can use the "cpu" command to verify which CPU is deconfigured. Enter "cpu X on" to reenable the CPU (X is the number of the deconfigured CPU).
Of cause it would make sense to call the next HP response Center and let them find out if this CPU better should be replaced ...

Best regards

Stefan

Re: CPU missing

Jeffrey F. Goldsmith — Thu, 13 Nov 2003 16:27:11 GMT

Hello, I just got done booting my server and when the server first started up I received my first error. It looks like I am having a hardware error.

I am attaching the information.

Re: CPU missing

Michael Steele_2 — Thu, 13 Nov 2003 18:05:42 GMT

Yes. CPU 3 is deconfigured.

Time for a replacement.