Rejecting I/O to offline device

[debolaz]

Just now, a proxmox server (Ie, not a vm) we've got running here stopped working. Looking at the screen output through the DRAC management card, I saw this message being flooded across it:

sd 0:1:0:0: rejecting I/O to offline device

Googling the error, it seems like the kernel for some reason removed a disk from the system. I have however no idea why, and I don't think it's a genuine hardware error. What I want to know though, is how I go about debugging this problem? Is there anything I can enable to capture the original error, etc? Or some magic switch I might try to prevent the problem from occuring?

 

[dietmar]

it seems like the kernel for some reason removed a disk from the system.

This usually indicates an serious hardware error.

 

[debolaz]

I'll try to enable remote logging of kernel messages and provide more details if it happens again.