Koozali.org: home of the SME Server
Contribs.org Forums => General Discussion => Topic started by: mazkot on July 27, 2009, 03:13:02 PM
-
Hi got a server running for 3 years now.
Just this morning it started to keep on restarting every hour or so have to shut it down.
What logs should I check to find out what caused the restarting or the problem?
thanks
-
Hi
I'd start to check mb/cpu temperature.. and ram too..
is the server connected to an ups?
Ciao
Stefano
-
room is cold enough. so temp is not a problem. it's connected to a ups. but lately we have some flactuations. maybe will run some memtest and cpu test.
what logs should i also check?
-
I too would look at the CPU cooling issues first, may not be logs to see but the physical box is worth looking at.
I'd have the case apart and lift out any dust, re-seat the RAM, Video Card (if it has one) and pinch up the motherboard mounting screws, try it then if no change remove clean re-thermal paste and replace the CPU fan assembly.
Just because the room is cold the fan could still be running slow or full of dust.
Very long shot...
I had one Supermicro server motherboard that would seemingly random reboot until I found the mounting points had oxidised and the Mobo grounding was sensitive and bad, quick loosen and pinch up mountings got it stable again. I know that is a real long shot but someone has to benefit from the hours I spent tracking that down ;-)
-
room is cold enough. so temp is not a problem. it's connected to a ups. but lately we have some flactuations.
you'd never be too sure.. :-)
random reboot is an hw issue..
anyway, I'd check /var/log/messages and dmesg first
Ciao
Stefano
-
Hi got a server running for 3 years now.
Just this morning it started to keep on restarting every hour or so have to shut it down.
What logs should I check to find out what caused the restarting or the problem?
thanks
It's a sign of faulty RAM.. if I were you I'll replace them first.
-
thomasch
It's a sign of faulty RAM.. if I were you I'll replace them first.
search here for memtest to give them a thorough workout.
More likely fix could be to remove the RAM chips and plug them in again (2 or 3 times). This will clean the contacts
-
Still in denial mode. Since have an identical machine started on the same day and same place bought. And is okay. It's hard to find a shop this day that would not give you a bad stock. can't say that the users of both machines have the same usage.
Thanks for the infos. Still a linux nood. to the important logs to check.
-
Still in denial mode. Since have an identical machine started on the same day and same place bought.
It's not an identical machine, since it's not rebooting constantly.
Abandon your denial and test the memory already. Or don't ask for our advice if you are determined to ignore it. :-)
Still a linux nood. to the important logs to check.
I can't work out exactly what you are saying there, but clearly you are asking for logs.
There may be no relevant logs. Faulty hardware can cause an immediate reboot, without there being any logs produced. Even if there are logs produced, you cannot rely on them being stored to disk before the reboot.
-
I'll add to the crowd...
A random reboot is a hardware issue. Start checking the hardware (ram, MB, cpu, dust free thermals, ups, etc) & forget the logs.
-
thanks. Running memtest no problem found. Just did some cleaning and blower. Seems dirt just creep in after 6 months of cleaning it. Might change the yearly to every 6 months of cleaning. It's been running fine for 3 days.
hope the problem is not in hiatus and would return after a week or two.
btw what's a nice cpu test program.
-
mazkot
Running memtest no problem found.
FYI running a short memtest cycle does not really test the memory out. You need to run memtest overnight, say for at least 12 hours, in order to fully stress and test the memory.