server hung

[Sac_3]

Hi All,

Need help in listing out the steps one would take to resolve a server hung issue. Being an admin what are all the steps you should check to identify the issue?

can someone list out the steps?

[Ross Minkov]

Sac,

To get you started in the right direction google the following -- sysrq, netdump, diskdump

you will find lots of complete HOWTO set these up and use them in various places, i.e. Red Hat has this doc for sysrq -- http://kbase.redhat.com/faq/FAQ_80_5559.shtm, netdump -- http://kbase.redhat.com/faq/FAQ_43_2467.shtm and http://kbase.redhat.com/faq/FAQ_43_4220.shtm, etc.

Also each one of these once installed on a Linux system will put some good documentation files (txt, README, etc.)

HTH,
Ross

[rmueller58]

can you access it at all?

Such as from a console?

If you can gain access via a console, then it is a matter of hung processes, such as SSHD, or other remote access process.

If you can get access from the console check to see if you have a NFS or CIFS mount point not responding.

I recently dealt an issue where CIFS mounts were failing on a particular flavor of Fedora, I had to YUM update fedora to overcome the issue.

[Zeev Schultz]

I recall looking into dump docs for Linux time ago,even putting netdump server to collect dumps (which were partial for some reason). Anyway,/var/log/messages (console / dmesg alternatively) is a first place to start and look for oops strings ,EIP etc.

Try this link if you indeed face oops:

http://kerneltrap.org/node/3648

Now,oops are a kernel panic not hardware failures which are discovered in various and not allways user friendly ways in Linux.