Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

I have several Ubuntu Server 8.04 machines at a remote location. Every couple of months or so, one of them would stop responding and need to be power cycled. From looking at my log files it seems that all my processes are running fine until at some point everything just stops.

I suspect it's a hardware problem, but I don't even know how to begin pinpointing the issue. Are there any diagnostics tools or techniques designed to track down these sort of problems?

I know this is a fairly general question, but I'm hoping for a general answer.

share|improve this question

5 Answers 5

up vote 2 down vote accepted

Hook up another machine and configure a serial console to get all of the kernel messages and such that come up. If it's a kernel panic or some other catastrophic problem, you'll see it there. Monitoring temperature and running a memtest are also recommended, especially if the console shows nothing abnormal before the wheels fall off.

share|improve this answer

I had a similar problem in the past, and it turned out to be heat related. Improving the circulation and adding a fan or two helped big-time.

Also, make sure you've got SMART enabled on your disks and have a look to see if maybe one of them is on its last legs.

You might want to install munin to monitor them all and see what's going on.

share|improve this answer
    
I had this problem, and it turned out to be the motherboard's chipset overheating. It had a giant heatsink already but was still overheating. After the damage was done, it didn't matter that I added extra cooling. I'm still not totally sure it was the chipset, but I swapped out everything else including CPU, memory, videocard, power supply, and hard drive. –  Joseph May 3 '09 at 17:36

Memtest would be the first point of call although if you can, ask the center to plug in a console next time it crashes. If the kernel is going, it should output something to screen.

share|improve this answer

Put in a comprehensive remote monitoring solution with something like Zabbix. Monitor aspects of system resource usage, as well as any hardware statistics that are available to the operating system (eg. fan speeds, temperatures and the like). That way, when your system next falls over, you'll have a number of data points you can look at to see what the problem is.

With this approach you may find, for example, that you have a process that goes out of control with RAM allocation, pushes the system into swap, and causes the out of memory killer to start carving its way through your running processes, leaving the machine unresponsive. Without monitoring, you couldn't have known that.

share|improve this answer

Too little information given to actually anything that would definetly work.

It would be good to know how you define "stops" responding ? Is it just the ssh that stops responding or some other service ? Any ideas if console is still responding ?

Any traces in the logfiles after the machine is back online after the reboot ?

Few options anyway to get you forward on gathering the information:

  • Enable getty on serial line and if you cant affort on buying a serial server, crosswire the serials between the machines. If one machine cant be reached via network, you can try to get access via serial.
  • install monitoring software and fetch status from lmsensors, smart tools tec.
  • send syslogs to remote machine.
share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.