[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: System Crash Troubleshooting

On Tuesday 28 February 2006 01:15, Tim Alberts wrote:
> Bazooka Joe wrote ..
> > I would think this is a hardware issue.
> >
> > start w/ replacing the ram and see if that fixes it.  if not keep
> > replacing components until stable.
> Well that's certainly a possibility.  However, every time in the past this
> happens, re-installing the operating system fixes it.  So while it's
> possible, I consider it unlikely and WAY TO EXPENSIVE to shotgun it that
> way.
> Point taken though, I think I'll run the RAM test program that is on the
> Fedora install disks.
> Unfortunately, I don't see how faulty hardware can cause the session
> messages that I keep getting.  Looking again at my logs it seems they run
> every five minutes which led me to check CRON.  the only thing that appears
> to be running every 5 minutes is mrtg and disabling that did not get rid of
> the messages.  It appears that some rogue program is running in the
> background and I need to narrow down what it is.



1/ You may not be using postgres .. but is it set up to initialise at system 
Check system-config-services.

2/Are you running syslinux on this system?  Check what mode you are running in 
(enforced/permissive/etc), and check the /var/log/audit/* logs for any 
messages which might be relevant.

3/Is this machine actually a DHCP server?  If not, make sure 
root# service dhcpd stop
root# chkconfig dhcpd off

4/ If you have no need for postgres on the system, remove the user postgres 
for the time being, and see what complains about a missing user.
(Edit /etc/passwd and comment out the postgres user line.)

5/ Check what services are enabled under xinetd.
root /etc/xinet.d/ # grep -HP "[E|e]nabled. =. yes" * 
(thats .<space>=.<space>).
(-H displays the file name for each match, -P treats the search string as a 
perl regexp )
Disable each service that uses xinetd one by one and see what happens.

6/ Did you ever play with configuring dhcpd or pam/ldap (or anything else?) to 
use postgres as the storage back-end?

7/ Do a find for all files owned by postgres.  Cross check that against the 
crond and xinetd entries.

8/  What's the content of your /etc/inittab file?  There's an entry in your 
logs that makes me wonder .. the line 
>Feb 14 17:11:27 msi2 init: Trying to re-exec init
suggests something is trying to respawn.  This might suggest that the Andy 
Green is on the right track, as if X has a problem it might well be trying to 
trigger a respawn.
Did you ever set up multiple screens?


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]