sam core dump

[Ceesjan van Hattum]

Hi,

It's been a long time speaking to you all..

I've got 2 (very) old HPUX 10.20 systems. I cloned them myself... but now all of a sudden one system gives me the error when i try to retrieve a userlist within sam:
UNEXPECTED EXIT: process LANG=C LC_ALL=C /usr/sam/lbin/samx -C -p 4787 -s users /usr/sam/lib/ug.ui exit with a non-zero exit status: Bus error(coredump).

Some advices might be to install a patch. BUT
i've 'diff'-ed the swlist -l product from
both systems and there is no difference.
Therefor i do not like patch at all.

Anybody?

[Chris Wilshaw]

On old systems, there used to be issues with a number of user-related commands (finger and last were affected if memory serves). The system would drop a core if there was a + or & character in the gecos field of the passwd file.

It's possible that this is the cause.

[Kent Ostby]

Make sure you are patched with SAM/ObAM patch.

There was a known problem where you needed to have at least PHCO_16344 on your box or later or you would run into this problem from time to time.

PHCO_22268 is the latest version of this patch.

Best regards,

Kent M. Ostby

[Fabio Ettore]

Hi,

problem is about some libs were corrupted (due to several reasons).
Advices to install or REinstall a patch (I think the SAM cumulative patch) should reinstall those libraries and should solve your problem.

Anyway before try:

- is the PHCO_15645 on systems? I had a problem about this patch and problem was solved by installing the latest SAM patch.

- just verify filesets are configured:
# swlist -l fileset -a state

if any filesets are not configured then
# swconfig \*


I hope this helps you.

Best regards,
Ettore

[Ceesjan van Hattum]

Hi,
Chris: I've check with:
for i in `awk 'BEGIN{FS=":"}{print $1}' /etc/passwd`
do
finger $i
done
It didn't give any exceptions nor errors.

Kent: PHCO_16344 nor PHCO_22268 are installed. The oldest system is running
without problems for a few years now.
The cloned system ran without this problem for about 1 year. So even IF 16344 was installed, i do not see HOW this problem can occur.

Ettore: some libs were corrupted. Is there a way how i can verify which libs it concerns?
All patches installed are configured.

For now, i refuse to install any patch.
Is there a way for a 'library verification'?

(points will be assigned later)
Regards,
Ceesjan

[Helen French]

Try 'swconfig \*' command. Also if you want to check the status:

# swlist -l fileset -a state

I would also check and compare the files under /etc/sam directory, since these files will cause 'unexpected exit' errors in SAM. Use /usr/newconfig directory files if you want to replace it or just copy from the other system. Also read this TKB document

TKB #A5627601:
http://www2.itrc.hp.com/service/cki/docDisplay.do?docLocale=en_US&docId=200000063228265

[Fabio Ettore]

Hi again,

it is possible to find corrupted libraries by using them! Maybe SAM is using any corrupted libraries and then core dump.
Anyway it is just a thought, since I said that a my similar problem was solved by installing the latest SAM patch. Of course, if you don't want to install patches then I hope that anybody else can suggest you other solutions.
Anyway

Best regards,
Ettore

[Ceesjan van Hattum]

swconfig \* showed interessting stuff, but nothing essential to my problem.
I've seen in SAM-logging the sequence on which sam failes:

Performing task "Count Users"
Performing task "Count Groups"
Performing task "Get Users"
Performing task "Count Groups"
Checking to see if kernel regen is required.
[..end..]

I only ask a 'list' of current users. It looks strange to me that a 'regen' (i.e. re-generation) check is needed at this point.. maybe some libs are corrupted indeed ?
Can anyone tell me why such a check is performed at this point?

[V. Nyga]

Hi Ceesjan,

I've also seen several times that SAM didn't run correctly from one day to the other.
I could never imagine why but lastest SAM patches always corrected it.

Volkmar

[Robert-Jan Goossens]

Hoi CeesJan,

Have you seen this doc ?

http://www4.itrc.hp.com/service/cki/docDisplay.do?docLocale=en_US&docId=200000065679432

Document description: sam UNEXPECTED EXIT
Document id: KBRC00002798

Regards,
Robert-Jan

[Ceesjan van Hattum]

Hi collegues,
Okay.. i have to admit that patching the system might solve my problem. Nevertheless, the other system (by disk-cloned) does not have the problem. Appearently one or a few bits might have fallen.. corrupting a file or something like it.
I will think about it... the supposed patches will invole a upgrade on both systems. ... i.e. upgrading a system that does not have a problem is not what i want
either.
The problem i have is only when i access the list of users inside of sam.. not groups or something else. Adding users can be done by hand and therefor i'll try to live with the problem for now.
THANKS to everybody who tried to help me.

Regards
Ceesjan