cimservd causing reboot

[BoMahony]

i have just added a second processort to a rx2620. Machine starts, registers both proc's and kicks off loading the HPUX Os that was installed.

However, when it gets down a bit, it throws up:
Aug 29 06:52:16 cimserverd[1227]: cimserver[1309] not running, attempting restart

And the machine powers off eventually.

Im going to try and start the machine without the new processor now, but is there anything else i can do? Is there anything i have done wrong? I havent made any changes to the OS.
(Im also quite new to HPUX)

[Steven E. Protter]

Shalom,

Could be a coincidence, e.g. hardware problem.

But:

take look at: /var/adm/syslog/syslog.log

SEP

[BoMahony]

Steven

i assume to do so would require me booting off cd, if the Os isnt coming up.

[Torsten.]

You should test the new CPU using the offline diagnostic CD after installation.

Mabe a contact problem or something else.

Regarding the cimserv: message says it is restarting the process, not the server!

[Torsten.]

CD download:

http://h20000.www2.hp.com/bizsupport/TechSupport/SoftwareDownloadEventHandler.jsp?redirectReason=SWD_CDROM_Request&swItem=ns-50516-1&prodSeriesId=447331&prodLine=8B&targetPage=http%3A%2F%2Fwww.hp.com%2Fcgi-bin%2Fbizsupport%2Fwgbu-eula1.pl%3FPath%3Dftp%3A%2F%2Fftp.hp.com%2Fpub%2Fsoftlib%2Fsoftware8%2FCOL4150%2Fns-50516-1%2FIPF0706.zip

or

ftp://ftp.hp.com/pub/softlib/software8/COL4150/ns-50516-1/IPF0706.zip

[Steven E. Protter]

Yes, boot off the cd/dvd

Boot

Interupt at the 10 second prompt. I think this is an EFI system. There is a menu system that should let you see the log files and find out why the system is crashing.

SEP

[BoMahony]

I booted up off the Diags cd, and while booting went for a smoke. By the time i had come back, the machine was off again.

So it looks like a HW issue. now the thing is , i logged into the MP, and checked the logs and i am getting:

170 BMC 2 0x2046D57080020BF0 020EA37002120300 Type-02 127003 1208323
29 Aug 2007 13:11:28
169 BMC 0x2046D57080020BE0 020EA87002120300 Type-02 127008 1208328
29 Aug 2007 13:11:28
168 SFW 0 *3 0x7680026F00E00BC0 FFFFFFFF00FFFF72 MEM_DIMM_THERMAL_LOAD_ORDER_WARN
29 Aug 2007 13:06:43
167 BMC 2 0x2046D56F61020BB0 FFFF0103FDC00300 Type-02 c00301 12583681
29 Aug 2007 13:06:41
166 SFW 0 0 0x168002C500E00B90 0000000000000000 BOOT_REBOOT

so there is a Thermal warning on one of the DIMMS. how do i find out which one it is? I obviously dont have time to load the Diags to check as the chances are the machine will turn off.

[BoMahony]

looks like the cimservd isnt causing anything. Looks like the amount of time the mail daemon is taking to start up is amounting to the server just turning off:

Start IPv6 router advertisement daemon ........................................... N/A
Starting mail daemon ............................................................. Aug 29 09:44:31 cimserverd[1183]: cimserver[1309] not running, attempting restart
Aug 29 09:44:32 cimserverd[1203]: cimserver[1309] not running, attempting restart
Shutting down machine, please wait.

Shutdown complete.
Aug 29 09:44:33 cimserverd[1242]: cimserver[1309] not running, attempting restart

GenericSysName [HP Release B.11.23] (see /etc/issue)
Console Login:
sync'ing disks (0 buffers to flush):
0 buffers not flushed
0 buffers still dirty
Synchronous Page I/O error occurred while paging to/from disk
device 0x00000000, 1K block #2348732, 0 bytes of 28672 bytes not paged, page addr =eff0031.c000000007114000, bflag =1f, bp = e00000012f07e340

Closing open logical volumes...
Done

[Torsten.]

Log on to the MP and run "PS", should look like this:

MP:CM> ps


PS
System Power state: On
Temperature : Normal


Power supplies | Fan
# State Type | States
-----------------------------------------------------------
1 Normal Type 0 | Normal
2 Normal Type 0 | Normal
...


Important:
Temperature : Normal

What do you get?

Any overtemp may cause the server to shutdown!

[whiteknight]

Bomahony,


Here is the few steps you need to do:-

1) Check if /etc/opt/wbem/cimserver_start.conf existed
when cimserver was not actually running,
#ps -ef | grep cimserver

2)remove /etc/opt/wbem/cimserver_start.conf
#rm /etc/opt/wbem/cimserver_start.conf

3) start cimserver
#cimserver

It should solve your problem

WK
please assign points

[Marcel Burggraeve · Accepted solution]

Might sound weird but did you put the plastic cpu cover back in place ?
Without the cover there might be some kind of weird airflow causing components to overheat.

[BoMahony]

Ok i was out for the last few weeks, due to injury. here is where i stand now.

The system had 4 256dimms from when we purchased it, and 4 1Gb Dimms i got recently. I removed th e 256s before i was out sick.

The system now stays up and doesnt reset. However it still hangs at the starting mail daemon.

I have booted off the diags cd and ran the diags on the HW in there - both procs and the 4x1Gb mem sticks.

Everything passes.

So where should i go from here to find out why the systme just hangs before the OS is usable?

If i boot up off an OS cd and check the local logs, what should i be looking for?

[Torsten.]

"it still hangs at the starting mail daemon."

I would think this is a network problem.

If the system is configured to use DNS, it will probably try to resolve the own name by using DNS - if there is no access to DNS (or the DNS doesn't know the server yet), it will look like hung for a while.

You can either wait or boot to single user mode and adjust the settings.

[BoMahony]

The port on the switch was dead. Go figure.

Anyways, I havent seen how far this time the load gets, as i am currently working from my deck, using network to get to the MP. However, once the MP starts loading the OS, i no longer get anything over the network - is there a reason for this?

Ill have to hook a monitor up to the console port tommorow to see how far its getting. Its still not coming all the way up though..

[Torsten.]

I have still the link open, here it is:

http://docs.hp.com/en/5991-4794/ch04s01.html#cihgcffj

Use the

Acpi(HWP0002,700)/Pci(1|1)/Uart(9600 N81)/VenMsg(Vt100)

entry (only) for input/output/error.

[BoMahony]

Weird - when i was downstairs, each Input/Output/Error, only had 4 entries - now they all have 8 - ill assume its because i am not connecting over the MP (i couldnt see these when connected via the console port)

[BoMahony]

This is crazy. I made the changes to the EFI shell using the network port. Restarted the machine this morning and it came all the way up to a login prompt, and can now be accessed via regular TCP also (without having to go via the MP)

[Torsten.]

Sounds like problem solved?

Consider to assign some points to everyone.

;-)