v2200 system panic

[June Cui]

Hi,
Our v2200 system crashed many times during the weekend after it has been up for about 30 days. Now the system won't boot anymore, even to the single user mode. Following is the message I got when the kernel is being loading:
interrupt type 18 (Data Memory Protection fault):
instruction address (pcsq.pcoq)=0x0.0x7f11878
instruction (iir)=0x4a730000(load/store)
target address (isr.ior)=0x0.0x000000f00081ff26
base register(gr19)=0x000000f00081ff2b
savestat ptr (ssp)=0x0.0x0000000007ffc3b0
savestate return pointer (ss_rp)=0x7f111e4
system panic:

If it's memory hardware problem, is it easy to take out the bad memory by myself on a v-class system?

I appreciate all your help!

June

[Bill McNAMARA_1]

Can you get the contents of the PIM hpmc?

Try boot with init S arguements.
ie BO -is

What is the KGM setting?

Later,
Bill

[John Poff]

June,

It looks and sounds like a memory problem. The V2200/V2250 boxes had some troubles with memory. The memory isn't hard to swap out, but if you have HP support on the box I'd suggest calling them and letting them fix it. On the V-class systems the bad memory gets deallocated and you have to run a GUI program from the test station/console to get it re-allocated after replacment.

Unless the box is unsupported or support is not an option for some reason, I'd call HP.

Good luck!

JP

[June Cui]

boot -is didn't work neither. The system panics when I booted to a single user mode.
Thanks
June

[Bill McNAMARA_1]

Can you cut/paste the power on self test of the system from the test station here.
That should identify memory probes.

You should open a support call in any case.

Bill

[MANOJ SRIVASTAVA]

Hi June

Please call HP ASAP . It looks like it is memory Problem , it is not difficult to change memory , but it will be better to call HP and get the MEMORY tested , changed .


Manoj Srivastava

[June Cui]

Bill,
Thank you for your help. Attached is the output of system boot message on the console.
I don't know how to use hpmc. Can you give me some instruction on this?

Thanks
June

[melvyn burnard]

A Data Memory Protection Fault is more often than not a hardware induced panic.
Enough people have recommended you contact HP, and I believe you now need to get the dump analysed and the hardware checked by HP, as this is not a small task. The longer you leave it, the longer it wiull take to get your system up and stable again.

[Bill McNAMARA_1 · Accepted solution]

it looks to me like the source of the problem is in Unix. The message
"panic: Data memory protection fault" points to a panic, but there is no
message of an HPMC.

I am sure, if you
reinstall the operating system and you patch it well, the problem will
disappear. A softer fix would be to analyse the dump for the real cause, but
this has to be done by an software expert.

Later,
Bill

[June Cui]

Bill,
Thank you for your help. Now I even couldn't get my console come up on the test station.

Thank you again!
June

[Ted M Johnson_1]

I haven't touched a V in a couple of years, but I noticed the nodemask is set on this box. Why would a V2200 (not a spp-ux X Class or V2500 ccNUMA box) have that set? Are we sure this isn't a botched firmware setting? A wrong fourth mode flag will cause strange things to happen.