dump device

[Sivasingam Santhakumar]

Hi,

I took over the administration of a L2000 server running HP 11.0. There is no crash dump device configured and previous admin has mounted a file system under /var/crash. I see the follwing error message /etc/rc.log
/dev/vg00/dump: error: device is fully occupied by a file system
EXIT CODE: 1

How could I fix this?.

Thanks

[Michael Tully]

You may find that the incorrect device/filesystem is defined either in the /sbin/init.d/savecrash or perhaps the mountpoint is incorrect. (see /etc/fstab) The standard is actually /var/adm/crash

[T G Manikandan]

check your /etc/rc.config.d/savecrash file.

change the SAVECRASH_DIR point to a different directory
The default is /var/adm/crash.

What is the output of
lvlnboot -v?


Revert

[Michael Steele_2]

When a crash dump occurs it first writes to the logical volume listed in lvlnboot and then out to disk with the 'savecrash' configuration. It can also be retrieved after reboot with the 'savecrash' command so all may not bo lost :

lvlnboot -v,

...to see if a dump has been configured with 'lvlnboot -d', stage one, and then:

savecrash -rf /tmp/file

...to still retrieve your crash dump and write it out to disk, stage two. You'll need a minimum of 400mb free, perhaps more, depends if you're saving more then just system data. (* application, process, et al. as well *)

I can't use the ITRC search engine right now to check on your error but it sounds like a pervious crash dump already preoccupies the reserved area and needs to be 'copied' into another file system area unless you've already got it on tape. If so then I'd remove it.

Check what /etc/rc.config.d/savecrash indicates.

[Michael Steele_2]

WAY TO GO MR. MICHAEL TULLY!

THE FIFTH MR. OLYMPIAN!!!!!!
THE FIFTH MR. OLYMPIAN!!!!!!
THE FIFTH MR. OLYMPIAN!!!!!!
THE FIFTH MR. OLYMPIAN!!!!!!