topic Re: Server down, Why? in Operating System - HP-UX

Server down, Why?

Sergio Tancredi — Mon, 23 Apr 2001 09:01:52 GMT

Hi All,
Last day my R-380 it went down and after reboot I 've found nothing that it could help me to understand why it went down.

I checked OLDsyslog, mail, shutdownlog whit no result.
/var/adm/crash is empty.

The only strange thing that I've noticed is that when I run swapinfo, no output regards memory it returns me.
It display only information about reserve memory and swap area.

So I have thought that the problem had to be an defective block of memory, but I have no errors using stm (Online Diagnostic).

Could you please help me???
Thanks in advance,
Sergio

Re: Server down, Why?

Paula J Frazer-Campbell — Mon, 23 Apr 2001 09:40:36 GMT

Hi
Have a look at /etc/rc.log and /etc/rc.log.old.

Also have a look at /var/tombstones/tsNN (pick last tombstone).

HTH

Paula

Re: Server down, Why?

Sergio Tancredi — Mon, 23 Apr 2001 09:55:25 GMT

Hi,
Unfortunately, I haven't in /var/ the directory tombstones because I've installed stm this morning.

I checked also rc.log and rc.log.old but this can't help me because I would understand why the server went down, and in rc.log I've found only messages of booting.

Do you have others help for me?
Thanks,
Sergio

Re: Server down, Why?

Barry O Flanagan — Mon, 23 Apr 2001 09:57:02 GMT

Might be worth checking who was last on the box with a :

last -10

And find out if anyone was running anything unusual or even if they rebooted the box.

Might be worth checking the /var/adm/cron/log also.

Re: Server down, Why?

Tim D Fulford — Mon, 23 Apr 2001 10:01:31 GMT

Try looking at /var/adm/shutdownlog
if the "uptime" & this file differs you may have a power problem! If not it will give a clue (it may be only a tiny clue) of why it was shutdown or failed!

cheers

Tim

Re: Server down, Why?

Bill McNAMARA_1 — Mon, 23 Apr 2001 10:08:20 GMT

have you any files in the lost and found dir.
This would suggest a power failure....
Do you have cleaning ladies that use power plug cleaning stuff....! not a joke!

Your shutdown log will tell you if there was a panic or a shutdown/ reboot issued manually.

Your old rc.log, look for fsck's of your root file systems indicating improper shutdown, typically power related.

Later,
Bill

Re: Server down, Why?

Sergio Tancredi — Mon, 23 Apr 2001 10:22:03 GMT

The date of shutdownlog is ... 29 FEB 2000.

It must have had the date of the last one reboot?

Re: Server down, Why?

Pedro Sousa — Mon, 23 Apr 2001 10:54:40 GMT

Hi!
Check /var/adm/ptydaemonlog.
does it have an output similar to:
Tue Mar 27 09:27:42 2001 pid=398: SHUTDOWN
Tue Mar 27 09:33:17 2001 pid=405: STARTUP
Tue Mar 27 09:33:18 2001 pid=405: INITIALIZATION COMPLETE

Also check:/var/adm/vtdaemonlog

If it doesn't have the shudown line, it's because the system didn't went down correctly.
So, on these conditions, I had to similar things hapenning:
- An UPS connected to the system was bringing it down.
- A user switched off my K thinking the key was to open the door :)

good luck.

Re: Server down, Why?

Sergio Tancredi — Mon, 23 Apr 2001 15:19:40 GMT

Hi All,

I haven't files in lost+found and I haven't the last shutdown line in /var/adm/ptydaemonlog .

...And I haven,t the file var/adm/vtdaemonlog.
I don't have UPS.

Now I haven't doubt that the system didn't went down correctly.

But now, why swapinfo don't show me the phisical memory???

Re: Server down, Why?

Curt Thompson — Mon, 23 Apr 2001 15:36:47 GMT

Hello Sergio,

An unexpected shutdown can be caused by defective system hardware. Regardless of whether STM was loaded or not, the HPMC (High Priority Machine Check) event will have been captured in non-volitile memory (PIM) for later analysis.

Now that you have STM installed, try rebooting the system again. If an HPMC exists in PIM, it will be copied to the ts99 file in /var/tombstones, at which point, you can post the chassis codes from that file.

Good Luck,
Curt

Re: Server down, Why?

Paula J Frazer-Campbell — Tue, 24 Apr 2001 06:29:22 GMT

Hi
As you have not traced as to why the server went down your next step is to take the server down and up under your control paying attention to all messages.

Why? well you have a server in an unknown state and swapinfo is giving you believe incorrect output.

HTH

Paula

Re: Server down, Why?

Luis Miguel Parra Chica — Tue, 24 Apr 2001 09:48:42 GMT

Try taking a look in the .history of your users. Another way to go down the system is killing the init process.

Good luck