Some time in the very early hours of Sunday morning, I woke to find my daughter knocking on the door to tell me that a server machine was making nasty popping noises and giving off nasty smells and smoke. I wasn't in much of a state do do anything other than disconnect it from the power and make sure nothing else was damaged. This is a server I'd been hoping to nurse along until RHEL 6 (or more accurately CentOS 6) is released, at which time I want to replace it with a much more modern and beefy machine.
I've been moving things off it, but among other things it still runs all my buildfarm members, and the engine for my git mirror. I would have moved all or most of that too, but I have encountered another problem: VMWare Server won't install nicely on modern kernels such as those used by Fedora 13, So I can't move those to my new temporary machine.
Anyway, after a little looking and sniffing, I decided that the problem was probably that the power supply had failed on the machine, and with any luck simply replacing it would fix the immediate problem. This could also explain the fact that there had been two spontaneous reboots without any symptoms while i was away, on a machine whose uptime has previously been measured in months. So I went out today and bought a new power supply, and after I installed it the machine came up just nicely. Whew! No wonder I hate dealing with hardware.
While I was cleaning things up, I noticed some old build products left from earlier problems that were preventing some of my buildfarm members from reporting on several branches, so I cleaned those up too and now I should see more regular builds from all of them.
I've been moving things off it, but among other things it still runs all my buildfarm members, and the engine for my git mirror. I would have moved all or most of that too, but I have encountered another problem: VMWare Server won't install nicely on modern kernels such as those used by Fedora 13, So I can't move those to my new temporary machine.
Anyway, after a little looking and sniffing, I decided that the problem was probably that the power supply had failed on the machine, and with any luck simply replacing it would fix the immediate problem. This could also explain the fact that there had been two spontaneous reboots without any symptoms while i was away, on a machine whose uptime has previously been measured in months. So I went out today and bought a new power supply, and after I installed it the machine came up just nicely. Whew! No wonder I hate dealing with hardware.
While I was cleaning things up, I noticed some old build products left from earlier problems that were preventing some of my buildfarm members from reporting on several branches, so I cleaned those up too and now I should see more regular builds from all of them.