We have a server (IBM x3850 8864) running Ubuntu 10.04 LTS. It's acting as a KVM host with several virtual machines on it (also running Ubuntu 10.04). The (physical) server sometimes crashes with error messages like
Checking monitoring graphs, there's no high load or any other unnormal thing going on before the crash. This kind of crash happens every 2-3 months. In our situation, it is too rare to just throw the box away, but too often to ignore the problem. So my questions are:
- What do these error messages mean?
- If it's not possible to tell from the error message what the problem is, what would be some next steps to investigate this issue further?