[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] xm shutdown timeout

  • To: John Levon <levon@xxxxxxxxxxxxxxxxx>
  • From: Keir Fraser <keir@xxxxxxxxxxxxx>
  • Date: Thu, 24 May 2007 14:44:52 +0100
  • Cc: xen-devel@xxxxxxxxxxxxxxxxxxx
  • Delivery-date: Thu, 24 May 2007 06:43:16 -0700
  • List-id: Xen developer discussion <xen-devel.lists.xensource.com>
  • Thread-index: AceeCbRK8qUeSwn8EdyL+wAX8io7RQ==
  • Thread-topic: [Xen-devel] xm shutdown timeout

On 24/5/07 14:45, "John Levon" <levon@xxxxxxxxxxxxxxxxx> wrote:

>> Perhaps pushing into state 'crashed' is the right thing to do, and treat
>> that state as a catch-all for exceptional termination conditions?
> I considered that. Whilst it's an improvement on what happens now, it
> still seems like the wrong thing to me: typically a crash will dump core
> and destroy, meaning I can't then log in and look around (or debug it
> live from the dom0). I imagine I could set 'preserve', but then I have
> to preserve real crashes too, which is not likely to be what I want.
> Since we can't have a new state, it seems like leaving it would be best.
> Perhaps renaming it to indicate it won't respond to a shutdown request?

Why would you want to dump core and destroy for one type of bug, but not for
another? Renaming the domain is an okay alternative.

 -- Keir

Xen-devel mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.