Re: Random Linux SERVER hangups (production)

On 2006-03-24, Ignoramus4723 <ignoramus4723@NOSPAM.4723.invalid> wrote:
> We have a linux server running FC3. Once in a while it would become
> bad. The symptoms would be that our main process is hanging and we
> also cannot logon via ssh.
>
> - At some point I may be able to log on as root, everything loks OK, I
> can do ps, uptime etc, but if I do tail -f <our-process-log-file>, the
> tail command hangs up in a bad way -- ^c or ^Z does not work. The file
> is a regular file.
>
> - After that, Ssh kind of works, connects up to the point of saying
> "Creating interactive session", and then it hangs. The system becomes
> unusable and needs to be rebooted.
>
> - No relevant messages appear in /var/log/.
>
> This is REALLY annoying and affects acceptance of linux as a server
> OS. So, I need some sensible ideas on how to track it.

It sounds like it might be some process (maybe your "main
process", whatever that is) is growing in memory footprint
until the system grinds to a very slow crawl. Running 'top'
in batch mode and 'vmstat 10' into a file would probably
provide some evidence of the cause.

--
Robert Riches
spamtrap42@verizon.net
(Yes, that is one of my email addresses.)