Problems with savecrash

[Jesper_2]

Hi everyone, I'm quite new to HP-Ux and now have I a problem that I have tried to find a solution for on docs.hp.com, but I have not find anyting.

My problems is that my server L-class "crash" I don't get any dump in /var/adm/crash to use when I try to solve the reason for the crash.

Would be grateful for quick help!!

[melvyn burnard]

what does lvlnboot say?
what is the output of crashconf?
what is /etc/rc.config.d/savecrash set to?

What if any errors/messages do you see?
what does /etc/shutdownlog say?
What OS are you on?

[Sajid_1]

First thing to check is the file - /etc/rc.config.d/savecrash

Check and make sure that the savecrash entry is not set to 0. Also check the save_directory name and find out if that's the default.

hth

[Bill McNAMARA_1]

crashconf will report configuration.

verify you have a crash/panic via reference to the /etc/shutdownlog

Post up your /etc/rc.log for verification also that the crash save process started.

Do you have space in /var/ ?
bdf

Later,
Bill

[Jesper_2]

The following in the /etc/rc.config.d/savcrash is set

SAVECRASH=1
SAVECRASH_DIR=/var/adm/crash


These two is the only that I have uncomented, the rest of the setting will be default because the are not set.

[Sajid_1]

hi,

check the file and it's permissions:
/sbin/init.d/savecrash

Check the log file and see if anything wrong - /etc/shutdownlog

Check the /etc/rc.log file for any other information

[Jesper_2]

Yes /var have enough space!

But I can se some strange thing in the rc.log that say :

Save system crash dump if needed
Output from "/sbin/rc1.d/S440savecrash start":
----------------------------
savecrash directory not set; defaulting to: /var/adm/crash
savecrash: Dump previously saved, use -r to resave

EXIT CODE: 2 - savecrash found no core dump to save
"/sbin/rc1.d/S440savecrash start" SKIPPED

Enable auxiliary swap space
Output from "/sbin/rc1.d/S500swap_start start":
----------------------------
swap device(s) active

_______________

[Sajid_1]

hello,

the message from rc.log is not an error. It's just a notification and is normal. It look like the system didn't have anything to dump, and that's why you didn't get a core dump.

hth

[melvyn burnard]

so when you say the system crashed, would you care to expand upon this?
exactly what happend?
and you may want to clear your message by isuing savecrash -r and then saving the dump, or try issuing savecrash -c to clear it.

[Sanjay_6]

Hi Jesper,

Do you have enough space in /var to store the crash. Check your /etc/rc.log file and see why the crash was not saved. If you don't have enough space under /var to save the crash files, you can save them at a different location. To do that change the default location by editing the file /etc/rc.config.d/savecrash and uncommenting the variable

# SAVECRASH_DIR=/var/adm/crash

and point it to a new location.

Then recreate the crash files, using,

savecrash -vr

Hope this helps.

Regds

[Bill McNAMARA_1]

I would imagine if /etc/shutdownlog reports a panic/crash that the only reason it didn't save on startup is that it didn't write to a defined dumpspace.

This is configured via lvlnboot.

post up the lvlnboot output

and set the dump definition via
lvlnboot -d /dev/vg00/lvol2 for example.

Later,
Bill

[Patrick Wessel]

Gentlemen,
Shouldn't we verify that the system crashed because of an OS panic before we urge it to write a dump?

Jesper,
Did you find any hint on the system console that points to a panic? Are there any entries in the GSP error logs with the alert level 12 (SW failure) or do you see in the GSP error logs legacy chassis code B800? Please take a look in the /var/tombstones/ts99 file. What is the timestamp of the last HPMC?
Maybe we find a hint what caused the hick-up

[Printaporn_1]

Hi,

try with
savecash -r
immediately after system startup ,

-r Resaves a dump that a previous invocation of savecrash has
marked as already saved. This is useful if the first
invocation did ran out of space, and enough space has since
been freed to try again.