HPE 9000 and HPE e3000 Servers
cancel
Showing results for 
Search instead for 
Did you mean: 

Crash dump analysis

 
manoj_pu
Regular Advisor

Crash dump analysis

Hi All,

One of our HP L3000 server rebooted due to some reason and I have analysed the crash dump. Below is the report of crash dump analysis.

Can anyone please help in concluding the exact reason of system reboot.

Thanks in Advance

Manoj


Created with WhatHappened: @(#) $Revision: 1.7 $ $Date: 01/01/22 11:11:28 $ for 11.i

For information on how to interprete the whathappened output, please see:
http://hprtdt04.grc.hp.com/FirstPass

System Name: HP-UX
Node Name: sterling
Release: 11.11
Version: U
Model: 9000/800/L3000-7x
Machine ID: 1102424661
Processors: 4
Architecture: PA-RISC 2.0
CPU is a: PCXW+
Physical Mem: 2097152 pages

This is a 64 Bit Kernel
The system had been up for 58.72 days (507298175 ticks).
Load averages: 0.05 0.03 0.03.

System went down at: Tue Aug 2 07:21:07 2005
+--------------------------------------------+
| Message Buffer |
+--------------------------------------------+

ull (1 block extent)
msgcnt 979 vxfs: mesg 001: vx_nospace - /dev/vg00/lvol_tmp file system full (1 block extent)
msgcnt 981 vxfs: mesg 001: vx_nospace - /dev/vg00/lvol_tmp file system full (1 block extent)
msgcnt 983 vxfs: mesg 001: vx_nospace - /dev/vg00/lvol_tmp file system full (1 block extent)
msgcnt 985 vxfs: mesg 001: vx_nospace - /dev/vg00/lvol_tmp file system full (1 block extent)
msgcnt 987 vxfs: mesg 001: vx_nospace - /dev/vg00/lvol_tmp file system full (1 block extent)

trap type 15, pcsq.pcoq = 0.38ca8, isr.ior = 0.fffffffffffffffc

linkstamp: Sat Jun 4 14:03:08 BST 2005
_release_version: @(#) $Revision: vmunix: vw: -proj selectors: CUPI80_BL2000_1108 -c 'Vw for CUPI80_BL2000_1108 build' -- cupi80_bl2000_1108 'CUPI80_BL2000_1108' Wed Nov 8 19:24:56 PST 2000 $
panic: check_caught_trap: Caught trap was not ours

PC-Offset Stack Trace (read across, top of stack is 1st):
0x00235a50 0x0028a8c0 0x00299078
0x0014e380 0x0019c9d8 0x000b1f00
0x0014dce8 0x00033f5c
End Of Stack
Leave with out tense and try best you gets result
5 REPLIES 5
Kent Ostby
Honored Contributor

Re: Crash dump analysis

Manoj ...

You'd need to post the output from :

trace event 0

(if you are following the using_q4 document, this would be in trace.out).

"Well, actually, she is a rocket scientist" -- Steve Martin in "Roxanne"
Kent Ostby
Honored Contributor

Re: Crash dump analysis

Also, do the following with the what.out:

grep LPMC what.out.

Looking at the length of your PC-Offset stack trace though, you should be able to get something from trace event 0 that is usuable.

If you can't, post the hptrace file generated when you run these commands:

echo "0x00235a50?i" | adb /stand/vmunix > hptrace
echo "0x0028a8c0?i" | adb /stand/vmunix >> hptrace
echo "0x00299078?i" | adb /stand/vmunix >> hptrace
echo "0x0014e380?i" | adb /stand/vmunix >> hptrace
echo "0x0019c9d8?i" | adb /stand/vmunix >> hptrace
echo "0x000b1f00?i" | adb /stand/vmunix >> hptrace
echo "0x0014dce8?i" | adb /stand/vmunix >> hptrace
echo "0x00033f5c?i" | adb /stand/vmunix >> hptrace
"Well, actually, she is a rocket scientist" -- Steve Martin in "Roxanne"
manoj_pu
Regular Advisor

Re: Crash dump analysis

Hi Kent,
trace event
-----------
q4> trace event 0
stack trace for event 0
crash event was a panic
panic+0x6c
check_caught_trap+0xa8
ulb_terror+0x0
copyin+0x158
CopyInIovec+0xa0
recvmsg+0x58
syscall+0x9a0
$syscallrtn+0x0

hptrace
-------
panic+0xA0: ADDIL L%0xffffb000,r27,r1
check_caught_trap+0xA8: LDD -160(r30),r31
ulbtraprtn+48: LDI 14,r28
copyin+158: B copyin+0x0094
CopyInIovec+0xA0: COPY r28,r3
recvmsg+58: ADDIL L%0xfff17800,r27,r1
syscall+9A0: B syscall+0x0630
$syscallrtn:
$syscallrtn: LDO -1232(r30),r19


There is no output from "grep LPMC what.out" command. I have created what.out file by doing "what " command.

Thanks & Regards
Manoj

Leave with out tense and try best you gets result
Kent Ostby
Honored Contributor

Re: Crash dump analysis

Manoj --

Generally the "ulb_terror" on the panic stack is some sort of hardware issue tripping up the software.

I'd suggest that you call HP to have the dump read to isolate where the problem is.

Best regards,

Oz
"Well, actually, she is a rocket scientist" -- Steve Martin in "Roxanne"
Joshua Scott
Honored Contributor

Re: Crash dump analysis

Do you have the latest patch bundles installed?

Check if this patch is installed:
PHKL_30954

swlist -l product | grep PHKL_30954

Josh
What are the chances...