system crash question: HP-UX 11.0

[Kirk Reindl]

My system crashed today. Do you know where I can find the crash information?

I have 2.4 GB of dump allocated.
And the physical system memory is: 2GB

I thought info was stored in /var/adm/crash, but maybe that is 10.2 only.

Thanks

[Uday_S_Ankolekar]

Hi,

You can check /var/adm/syslog/syslog.log file and /var/adm/syslog/OLDsyslog file for possible cause for the crash.

/var/adm/crash where crash dump normaly stores. But you need q4 to analyze it.

Goodluck,
-USA..

[Kirk Reindl]

What I think is strange is that I don't even have a /var/adm/crash directory on this system in question. Do I have to make the crash directory in order for the OS to dump the crash file???

[Roger Baptiste · Accepted solution]

<>

Check /etc/rc.config.d/savecrash file and see whether
SAVECRASH=1 and
SAVECRASH_DIR=/var/adm/crash
(or any other directory
which exists and is large)

If this is set correctly
and you do not have /var/adm/crash, obviously
savecrash would fail.
To confirm this, check
/etc/rc.log file which will
show whether it even attempted
to save crash.

You can still attempt
to save the dump, by doing:

# crashconf -v |grep dump
-> mulitply this by 4
and that is the approximate size of dump
in Mb , you would need.
It is dynamic and depends
on the state of the system

#create a filesystem with
that size

#mount and cd to the directory

#savecrash -r .
-> this will attempt to
save the dump in the
current directory
(provided it is not
overwritten by swap)

#once you get the dump,
you can run q4 analysis
or let HP support do it
and debug the info.

HTH
raj

[Uday_S_Ankolekar]

Hi, Again..

How is your /etc/rc.config.d/savecrash is set.
There will be savecrash=n and savcrashdir=
Is it set to 1 nad /var/adm/crash respectively..?

On OS 11 you can check tombstone file .The latest file will be /var/tombstones/ts99

GoodLuck,
USA..

[ramesh_6]

The configuration file for crashdump is /etc/rc.config.d/savecrash

Check the following parameters in the file

SAVECRASH which should be set to 1 abd
SAVECRASH_DIR=/var/adm/crash
{ or any other directory or filesystem
which exists }

If the above parameters are not set perfectly your savecrash will fail definitely.

The /etc/rc.log file will show you it attempted to save the crash.

Ramesh

[Bill Hassell]

HP-UX has always shipped without the required /var/adm/crash directory, and therfore, the crash dump is not saved on disk. If you work quickly after a reboot, you can manually run savecrash command. You can even save it to tape rather than use up disk space.

I say 'work quickly' because the typical; setup is to share the dump area with primary swap space, so the dump remains intact until the first page of swap is written--and then its gone.