1 CPU Failed

[Vipin Singhal]

Hello friends,
I have 8 CPUs in my Server but now it is showing that only 7 CPUs are working.

these are my server details:
#machinfo
CPU info:
Number of CPUs = 7
Platform info:
model string = "ia64 hp server rx7620"
OS info:
sysname = HP-UX
release = B.11.23

Please suggest what should i do to make my 8th CPU up.
Thanks
Vipin

[Pete Randall]

I believe you will have to place a support call with HP to get this fixed.


Pete

[Andy Torres]

Verify whether it has failed or not first. Any messages in your syslog? Maybe you could check your system events in the GSP.

[HGN]

Hi

You can check the logs on the console and also place a call with HP ,they will direct to send them the logs and based on that they would suggest what needs to be done.

Rgds

HGN

[Vipin Singhal]

This CPU got failed a long time before and I am here new joinee in my company.

How will i verify whether it has failed or not?

[Andy Torres]

Try running ioscan from the commanline to verify your processors, and their state:

ioscan -fnCprocessor

You should see them all CLAIMED, at least.

You can also try running stm, or cstm. Check the man pages if you're unfamiliar with stm/cstm.

[Vipin Singhal]

# ioscan -fnC processor
Class I H/W Path Driver S/W State H/W Type Description
===================================================================
processor 0 0/121 processor CLAIMED PROCESSOR Processor
processor 1 0/122 processor CLAIMED PROCESSOR Processor
processor 2 0/123 processor CLAIMED PROCESSOR Processor
processor 3 1/120 processor CLAIMED PROCESSOR Processor
processor 4 1/121 processor CLAIMED PROCESSOR Processor
processor 5 1/122 processor CLAIMED PROCESSOR Processor
processor 6 1/123 processor CLAIMED PROCESSOR Processor

It shows only 7 CPUs while I have 8 CPUs

[Bill Hassell]

The 8th processor has failed (ioscan cannot make it communicate) and must be replaced. You'll need to call HP to replace it.

[Lee Harris_5]

It might be worth checking the chassis logs of the server via the GSP console to see if that gives any more info.

If you don't have GSP access you can do it via CSTM using the command ru logtool then ce - this will let you view chassis errors.

Also, doing a sel path system followed by il in CSTM might also give you more info about the failure of the CPU to pass on to HP support.

HTH

[Andy Torres]

Yup. Definitely a hardware call. Not much else to check at this point. Good luck.

[Vipin Singhal]

hey Lee,
how to use ctm command?
when i use:

#cstm
cstm>

then i type "sel"
cstm>sel
^-- Syntax error. --
Expected the selection keyword "device", "path" or "all".

Please suggest how to use cstm command

[HGN]

Hi

Here is the cstm syntax

#cstm
cstm>map
cstm>sel dev 1 (select CPU 1 is an example)
cstm>info
cstm>il

Hope this helps

Rgds

HGN

[Andy Torres · Accepted solution]

Try running this as root:

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

[Vipin Singhal]

hi
when i was trying this command:
echo "selclass qualifier cpu;info;wait;infolog" | cstm

^-- (Information) is currently disabled. --

^-- (InfoLog) is currently disabled. --

[Andy Torres]

Try running:

/sbin/init.d/diagnostic stop
then
/sbin/init.d/diagnostic start

Then try this again:

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

Maybe that'll help getting STM info.

Hopefully, you've opened your hardware call with HP by now. Don't think you can fix this on your own.

[Vipin Singhal]

/sbin/init.d/diagnostic stop
then
/sbin/init.d/diagnostic start

will it effect to any other running problem?
may i run these command now

My server database is up

[Andy Torres]

Shouldn't be a disruption. I stopped and started mine to be sure.

[rmueller58]

GSP from the Console, cstm from SSH or Telnet. check the GSP logs.. I'd call HP for Hdwe support, have them replace it.

[Vipin Singhal]

ok, let me try

[Devender Khatana]

Hi,

This sort of situation normally comes when a CPU is disabled by system during reboot followed by some LPMC errors. But this is surely reported in syslog.log.

Does your system has I-COD installed? If yes and you have system installed with 7 CPU licenses then you can check CPU status using "icod_stat" command.

#icod_stat

HTH,
Devender

[Andrew Merritt_2]

There should be an event logged in /var/opt/resmon/log/event.log if the diags are disabling the CPU, so take a look there.

What version number does STM show when you run it?

Regarding the STM CPU info tool, on 11.23 the CPU information is shown by the 'system' info tool, there is no separate CPU Info tool (which was the case on 11.11), so you need to do:

echo "selclass qualifier system;info;wait;infolog" | cstm

Andrew