Server hangs without any message

[mjsunil]

Hi,
Our Redhat linux As 2.1 server hangs without any message. No indications at the hardware level. Is there any option where I can enable the debugging so that we can trace the root of the problem.
Any suggestions.

Rgds - mjs

[Steven E. Protter]

Welcome to the club. Happened to one other user this week on the exact same OS.

What kind of server, RAM, etc does it have?

You can enhance syslogd logging level in the syslog.conf file

You can break out kernel into its own file which could be useful since the firewall runs on the kernel.

SEP

[mjsunil]

Hi,
We have Dell Power edge and I have enable in syslog.conf to log all messages as
*.* /var/log/allmessages
Is there any thing else to enable?
rgds - mjs

[mjsunil]

I have hyper threading enabled and I will try disabling it
rgds - mjs

[Andrea Rossi]

same problem. we have tried to
- upgrade to rh 3update 1;
- upgrade bios
- pull off both cpu
- pull off memory
the problem persists.

[Anu Mathew]

What kernel verion do you run? Just curious there. Have access to RHN? How about patch compliance?

You can write a script to monitor the load ave of the machine, collect detailed info on processes running.

Need more to tell you more.

--AM

[Vitaly Karasik_1]

- what is your server model?
- your patch level? [update 5 is current]
- do you have 3rd party kernel modules?
- http://www.linuxdevcenter.com/pub/a/linux/2001/11/01/postmortem.html Linux System Failure Post-Mortem
- netconsole/netdump [included in RHEL] may help you (I haven't tried it)

[Andrea Rossi]

diagnostics:
A - put in /etc/syslog.conf
kern.* /dev/console
*.emerg /dev/console
B - run "top d1" on the console

tip:
just disable the hpasm.

[xyko_1]

Hi,

if your server simply hangs, with no message at all, my guess is temperature.

Kernet 2.4.x does not report that kind of problem, 2.6.0 does notify.

Try to verify if your server isn't getting hot.

Another possibility is a defective memory card or slot (most possible the card).

But begins with the temperature and if this is your problem please notify us.

Regards,
Xyko

[HGN]

Hi

The server hand can be caused both by hardware and software which is running on it. Not sure what kind of hardware you are running if you are using a Compaq/HP server you can load Compaq Insight Manager and set the snmp to send traps and on the Compaq Insight Manager Console(this needs to run on a different server)you can go and check for any hardware errors, this will report cpu./memory/fan issues on the server which can cause the server to hang.
For software crash even the messages file would not say anything, you can enable for getting a core dump on another server, If you are under RedHat support they would suggest this and also they would recommend you to upgrade the kernel,we ahve seen this happen many times in our environment.

Rgds

Gopi