Hi Team,

I am facing problem of Linux server going to hang state.
It is very random and also very rare, during this time I am not able to login to server.

I dont see any output on console screen.

Also when I see /var/log/messages I dont see any output during the time server was hanged.

I doubt that there is some process which is taking all Server resources and server is hanged. That is why no logs are seen in messages file as well.

PLease suggest how I can verify this. Also how can I find the culprit process.

Regards
Dhiraj

If there is nothing in the /var/log/* files then chances are it is a hardware problem. How does the server get out of the hung state?

Hi

Server has to be rebooted. Also checked with HP. And as per them there is no hardware issue.

I've had similar issue on Solaris cluster with Oracle, and was ping-ponged from SUN to Oracle

After this happened I created script that gave me ps and sar for each 5 minutes on the system and cron to put those data into file.
Then created other cron to rotate it on daily basis, so I have all data during one day in one file.

when the issue happened again, I could see what process ID had most of the resources and found out that Oracle was the one who crashed one node of the SUN cluster

you could try something like that

Thanks soulman_yu, a good suggestion

Can you tell us how Oracle software was crashing a node? AFAIK (theory!) it should not be possible for applications running in user space to halt the kernel ... ?

<duplicate post deleted>