There is something strange going on with my colo box. I tried to reboot it last month and it didn’t come up – I had to call my provider and get them to power cycle it. Nothing useful in the logs.
Yesterday I had to install a security update to the xen hypervisor, but I didn’t reboot. This morning, I discovered that the websites working on the xen guest (the domU in xen parlance) were not working. So I tried to log in, or ping, and discovered it wasn’t talking to the network. Fortunately the xen host (aka dom0) was working – I could log into it, then use xm console xen1 to log into the guest. Couldn’t find anything wrong, except it’s not talking to the network. Even “ifdown eth0; ifup eth0” doesn’t cure it. So I tried to reboot the guest, but it didn’t seem to come back up. I wondered if the hypervisor update I installed yesterday was the problem, so then I rebooted the whole computer, and it didn’t come back up either.
I drove down to the colo facility, and connected a monitor and keyboard, but nothing showed up. On the front panel, there are a couple of blinking lights. I power cycled. It came up just fine. Logged into the host, xm consoled into the guest, verified that I could ssh out, and from my home computer I could wget a few web pages from it. Issued a reboot command, and it booted just fine. Poked around the BIOS settings to see if there was something about not booting if there wasn’t a keyboard or something stupid like that, but couldn’t find anything. Booted, verified once more, and came home.
Until the next time, I guess.