atlanta34 required an emergency reboot. The Xen stack on this host was updated and Linode are booting up now.

-Tom

My linode on Atlanta34, which had run perfectly in the couple years since I acquired it, has had all of its only 3 crashes in the last few months.

Each time the linode becomes completely inaccessible, the processor monitor in linode manager pegs at 100%, and the system messages log gets an entry like this:

Jul 21 16:01:56 siouxsports kernel: BUG: unable to handle kernel paging request at virtual address 2b6611ec
Jul 21 16:01:56 siouxsports kernel:  printing eip:
Jul 21 16:01:56 siouxsports kernel: c04c1860

(I have the rest of the most recent one, only, if it's useful to anyone. The kernel is 2.6.18.8-linode16 #1 and the process is Apache httpd).

This time, it had been down for a couple hours before I noticed and issued a restart. The restart seemed to hang for about 5 minutes in the shutdown phase, which had never happened before. In what I suspect isn't a coincidence -- 5 or 10 minutes after I attempted the shutdown that hung, you came in and rebooted the host.

The delay in restarting is the linode manager attempting to do a clean shutdown, but this fails (machine has crashed!) and timing out and then doing a hard reboot.

I had plenty of BUG crashes with 2.6.18.8. I've switched to 2.6.29-linode18 and my machine has been a lot more stable.

_________________
Rgds
Stephen
(Linux user since kernel version 0.11)

The weird thing to me is that the machine seems to have crashed exactly when I tried to reboot my linode. My linode, itself, had been down for about 2 hours with linode manager statistics of 100% CPU for the entire time, so it was reporting. Only when I asked the linode manager to reboot my linode did the entire machine become unresponsive and tasaro come in to restart it. Possibly a coincidence, but seemed worth mentioning.


Thanks for the tip. Hoping for a "set it and forget it" choice, I've just always left it on Latest 2.6 series.