From mboxrd@z Thu Jan 1 00:00:00 1970 From: Keir Fraser Subject: Re: Re: oxenstored in stubdom ? Date: Sun, 22 Aug 2010 10:44:51 +0100 Message-ID: References: <4C70EA62.8000807@eu.citrix.com> Mime-Version: 1.0 Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <4C70EA62.8000807@eu.citrix.com> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xensource.com Errors-To: xen-devel-bounces@lists.xensource.com To: Vincent Hanquez Cc: =?ISO-8859-2?B?o3VrYXN6?= =?ISO-8859-2?B?IE9sZbY=?= , "xen-devel@lists.xensource.com" List-Id: xen-devel@lists.xenproject.org On 22/08/2010 10:14, "Vincent Hanquez" wrote: >>> oxenstored is already restartable (or used to be and easy to fix if it >>> was broken), so from a xenstore point of view, you could already restart >>> dom0; Obviously this would block all the domains that try to do a >>> xenstore query, but if the dom0 is restarted quickly enough this >>> shouldn't be too noticeable since a normal working domain shouldn't use >>> much xenstore after starting up. >> >> So that's "very probably restartable" then? ;-) > > well yes, "very probably" is pretty good odds i think. :p > > more seriously, it depends from which perspective you're looking at the > dom0 restart problem. But according to previous experience during > oxenstored development, i'm pretty sure that oxenstored would cope and > that most of the problems are elsewhere in the stack. moving oxenstored > to a stubdomain is almost orthogonal (roughly 89 degrees.) I don't think xenstored-in-stubdomain is the big barrier to dom0 restartability, that's for sure. Personally, I don't think full dom0 restartability, for things like seamless dom0 kernel upgrade, will ever be achieved. But I think particular vulnerable or critical services within dom0 can be made restartable. -- keir