system crash analysis

[Usman]

our checkpoint firewall server running HP-UX 11.0 crashed without giving errors in syslog. i have analyzed the crash dump and below is the stack trace for event 0 from q4.

-- trace event 0
q4> trace event 0
stack trace for event 0
crash event was a panic
panic+0x14
report_trap_or_int_and_panic+0x4c
interrupt+0x1e8
$ihndlr_rtn+0x0
ire_walk1+0x84
ire_walk+0x94
ill_fastpath_ack+0x5c
ip_rput+0x4f8
putnext+0xc4
fwstrmodrput+0xac
fwsm_rput+0x10
csq_turnover_with_lock+0x7c
putnext+0x244
hp_dlpi_mblk_fast_in+0x7c
hp_dlpi_mblk_intr_put+0xc0
str_spu_sw_isr+0x1a4
sw_service+0x8c
mp_ext_interrupt+0x108
ivti_patch_to_nop3+0x0
idle+0x444
swidle_exit+0x0


i have also attached the output from the command
echo "msgbuf+8/s" | adb -m vmunix . > /tmp/msgbuf

any help in pointing to the cause of this crash is more than welcome.

Thanks

[U.SivaKumar_2]

hi,
page fault occurs when a process tries to access memory that is out of it's limit or
the memory that is not present. A software
can cause this if it tries to acccess the
system memory.
As a matter of routine , check your memory usage and swap usage , log a call to HP and
get the hardware certified.

regards,
U.SivaKumar

[Stefan Farrelly]

Its simply a data page fault - basically the kernel panic'ed and crashed when it shouldnt have. Simply a patch issue. Your patches arent up to date. Load the latest patch bundle (March2002) or if you just want some individual patches go to the itrc home page, select individual patches, do a search for 'data page fault' for your HP-UX version and it will list the patches you should install to fix it.

We had the same problem here recently - March2002 patch bundle install fixed it.

[S.K. Chan]

First of all I think only HP response center would know how to direct this stack trace to an expert to translate what it meant. It's not an easy thing to understand. At the end of the day it is usually going to be patch installation that will solve your problem. I have ha similar crash, stack trace not quite the same but the "attachment" looks very similar to mine when it happened. I was told to install ..
PHNE_21038 cumulative ARPA transport patch
which should resolve the "race conditions" that are seen in the stack trace. And that I did and I never had the crash again.

[Trond Haugen]

I would advise to contact your local Response Center which would most likely like to have the deum-files sendt in for analyzing.

Regards,
Trond

[Clemens van Everdingen]

Hi,

I would check as well as HP as the guys from checkpoint FW.
If no contract I would suggest you started with installing the latest Arpa and streams patches and the dependencies.

C.

[melvyn burnard]

I believe this is fixed in PHNE_19375, superseded by PHNE_26771.
Do not forget the dependencies!