HP-UNIX was in a hung state

[djoshi]

Hi All,

Today morning our HP 9000, running HP-UNIX 10.01 was in a hung state.Display panel was showing FLT FBE/FB0, neither the console nor network was working. We hard boot the system and now the system is up.

Any one having idea about the problem? please advice me how to troubleshoot this, we do not have any support from HP on these UNIX boxes.

Thx,

[Prashantj · Accepted solution]

Hi Joshi,

Please check the following path.
#/var/adm/syslog/syslog.log
#/var/adm/syslog/OLDsyslog.log
#/var/adm/crash
#/var/adm/vtdaemonlog
#/var/adm/ptydaemonlog
#/etc/shutdownlog
#/var/tombstones/ts99 (Check the timestamp column)

Hope this will help

Prashant

[Bijeesh]

Hi,
It may be a hardware issue.Check the syslog Old syslog and /var/adm/crash files.Is there any extra h/w like n/w interface card, remove it.
can you mention the exact model number?

[Torsten.]

What server model?

# model

[djoshi]

The exact model is Hp 9000/800/K410

Thx,

[djoshi]

Is anyone having any idea about error code why the system was hung?

Secondly, if the problem arises again and system console and network will hung, what should we do to troubleshoot.

Thx,

[Patrick Wallek]

According to the K410 manual at:
http://ftp.parisc-linux.org/docs/platforms/A2375-90004.pdf

The FxxF codes are normal runtime codes.

However, the FLT is definitely a FAULT which is not normal.

The codes on the K-series server displays are usually 4 character codes. The FBE / FB0 seem to be incomplete codes.

Chapter 4 from the link above lists the various front panel codes for the server.

As far as troubleshooting goes if/when this occurrs again, make sure you note down the EXACT codes on the display and any messages that might appear on the console. If both network and console are hung, you may be out of luck as far as troubleshooting goes.

[djoshi]

Thx, all for suggestion.

Yesterday I change only the crontab and add below script. Does anyone think that system hang problem could be this reason?

0,5,10,15,20,25,30,35,40,45,50,55 * * * * /informix/home/wipe_log.sh > /tmp/clearLog.txt 2>&1


#!/bin/sh

date
bdf /informix
ls -l /informix/xyz.log.log
cat /dev/null > /informix/xyz.log
date

[Dennis Handly]

>Does anyone think that system hang problem could be this reason?

I wouldn't think that simple script could cause that crash.
Any reason you are using 10.01 vs 10.20?
Or a supported 11.11?

[djoshi]

Thx Denny,

The reason behind we are running 10.01 because our legacy applications, 4gl and informix runs only on 10.01.

Thx,