I apologize for the short notice, but I feel this is important enough to push out as soon as possible. I believe I have a work around for the bug in Xen that caused dallas38's host filesystem to become corrupted. I've done some testing and am unable to reproduce the problem. Since this bug exists on all our Xen hosts, it's only a matter of time before they too are affected by this problem.

Host dallas38 will be rebooted 3:00 PM Eastern (approx 1 hour).

Your Linode will be gracefully shut down, the host will be rebooted, and then your Linodes will be rebooted. I expect 10-15 minutes of downtime.

If all goes well, I'll be scheduling a reboot for the other Xen hosts, and then the pending Xen-beta sign ups will be processed.

-Chris

Shutting down Linodes on dallas38...

Reboot complete.

-Chris

This was only for the host filesystem corruption bug, not for the client IO lockup, is that correct? Just checking.

Yeah -- I expect this to only fix the host fs corruption bug. I guess there's a slight chance it will fix the IO hang, but it still points to a problem in the frontend blk driver inside the guest kernel.

-Chris