crash during untar of files

[David Bacon_4]

Hi,

Wondering if anybody has seen any patches or has experienced problems while
un-tarring files. I seem to cause my machine to crash every once in a while
when un-tarring files that I am moving from another HP machine (save version).

Here's the details:
HP/UX 11.00 32 with mirror/ux and online jfs
When un-tarring files, the untar goes along okay then hangs and the machine
crashes.
what I have in shutdownlog is:

16:23 Tue Mar 14 2000. Reboot after panic: , isr.ior =
0'10240024.0'46d1a128
and

11:41 Mon Feb 28 2000. Reboot after panic: Data page fault

It does happen regularly and I can generally untar the file after the machine
comes up.

Thanks in advance for any pointers.
david bacon

[Melvyn Burnard_1]

Just as a quick check, the two message syou show from /etc/shutdownlog show 2
separate types of panic.
The one from 14th March is the one that I believe you may want to concentrate
on first, as the panic string is normally logged in the shutdown log, like Data
Page Fault, but in this instance their is no panic string.
This is normally indicative of either an HPMC, or hardware crash, or a TOC due
to the TOC switch/button/command being pressed, or a TOC caused by
MC/ServiceGuard.
Did the system save the panic in /var/adm/crash? If so, there should be a crash
directory for the date and time of the panic, and in there wil be an INDEX
file. What does this say?

I believe the best suggestion here would be to get the dump analysed by your HP
Response Centre to see if they can pinpoint the cause of these crashes, as they
may not be related.

[David Bacon_4]

Yes I did get a partial panic, unfortunately I did not have suffecient space
available in var to accomodate the full dump.

I think that I will probably have to wait for the next time to get the full
analysis done.

thanks,
david bacon

----
the contents of the INDEX are:

comment savecrash crash dump INDEX file
version 2
hostname unixh
modelname 9000/800/K380
panic , isr.ior = 0'10240024.0'46d1a128
dumptime 953067999 Tue Mar 14 16:06:39 EST 2000
savetime 953068972 Tue Mar 14 16:22:52 EST 2000
release @(#)B2352B/9245XB HP-UX (B.11.00) #1: Thu Nov 6 01:58:21 PST 1997

memsize 4026515456
chunksize 67108864
module /stand/vmunix vmunix 11439240 708921907
error savecrash: Insufficient space to save full core dump

error savecrash: Could not completely process dump image area 1

[Melvyn Burnard_1]

OK David, for this panic in particular, you either had an MC/ServiceGuard TOC,
(are you running MC/ServiceGuard?), someone issued a TOC command, or you have
had an HPMC.
See if you have a directory called /var/tombstones. If so, there should be
files called ts?? where ?? is a number.
Look at the latest one there, and then vi it, or even better do:
grep -i timestamp /var/tombstones/ts99 (as an example)
If you get something like the following:

Timestamp = Fri Aug 15 15:50:15 GMT 1997 (19:97:08:15:15:50:15)
Timestamp = Fri Aug 15 15:50:07 GMT 1997 (19:97:08:15:15:50:07)
Then the timestamp is not relevant. If however you get a timestamp that matches
the panic time, you probably have had an HPMC, and would need to get a hardware
engineer to have a look at these.
Hope that helps

[David Bacon_4]

Hi,

No I am not running MC/SG.

I checked out /var/tombstones, there are no matching timestamps with the panic
time.

thanks,
david bacon