Fedora Core Freezes

baddah · 08-31-2007, 07:00 AM

Hi,

i've got a server that froze last week friday at 2pm.I restarted the server and went through the logs,could not find anything.It's been runnig fine for a week now and it froze again today(friday at 1pm).I though it might be a cron,but theirs no cron scheduled to run.Their is nou hourly crons and all the other ones runs at 4am in the mornings.

This is a part from my /var/log/messages

for last week

Code:

Aug 24 14:01:01 st2 crond(pam_unix)[23436]: session opened for user root by (uid=0)
Aug 24 14:01:02 st2 crond(pam_unix)[23436]: session closed for user root
Aug 24 15:51:43 st2 syslogd 1.4.1: restart.

and for the freeze now..

Code:

Aug 31 12:01:01 st2 crond(pam_unix)[3456]: session opened for user root by (uid=0)
Aug 31 12:01:01 st2 crond(pam_unix)[3456]: session closed for user root
Aug 31 13:01:01 st2 crond(pam_unix)[4359]: session opened for user root by (uid=0)
Aug 31 13:01:01 st2 crond(pam_unix)[4359]: session closed for user root
Aug 31 13:42:02 st2 syslogd 1.4.1: restart.

What can it be?Does the syslogd 1.4.1: restart.(which is the time we restarted the server) mean that it was still up,although i could not ping it,and their was no activity?I'm really confused.This is a new server,and it never froze during testing.

Any suggestions would really be appreciated.

Thanks

macemoneta · 08-31-2007, 11:04 AM

It could be many things.

The most common is a thermal problem - the temperature of a component gets too high. Check that your fans are all working properly, and monitor the CPU and motherboard temperature via lm_sensors.

It could also be caused by a misbehaving bus card. I had a USB/IEEE1394 card that worked fine - until backups to an external drive connected to it kicked in overnight. The system froze, because under heavy load the card would fail while holding the bus.

It could be a power problem. Are there any large power load devices on the same circuit as the server? A laser printer is a common machine killer, because of the high startup power demand.

baddah · 09-02-2007, 09:05 AM

Hi, Thanks for the reply

Quote:

The most common is a thermal problem - the temperature of a component gets too high. Check that your fans are all working properly, and monitor the CPU and motherboard temperature via lm_sensors.

I have installed lm-sensors and will monitor this to check.Funny thing though,this is the second server we have on site thats giving problems,so i dont hink its a CPU or fans issue.Both servers had the same PCI PRI Card(Its a Asterisk Voice Server),so maybe the card is giving problems.Any logs i can check maybe that will list possible PCI problems.Asterisk's logs seems fine?

The funny thing is that it happend on the same day a week apart,but i've checked everywhere for crons,etc and could not find any one that is scheduled to run at that time.I guess the only thing really is to replace the PCI card and check then.

Thanks