[Openstack] this can crash devstack

Peter Cheung mcheung63 at hotmail.com
Fri Jan 3 17:53:36 UTC 2014


Thanks Joshua
from Peter

> From: harlowja at yahoo-inc.com
> To: sean at dague.net; dtroyer at gmail.com; mcheung63 at hotmail.com
> CC: openstack at lists.openstack.org
> Subject: Re: [Openstack] this can crash devstack
> Date: Thu, 2 Jan 2014 18:01:15 +0000
> 
> If u want to try an alternate: http://anvil.readthedocs.org/
> 
> It goes through various stages:
> 
> 1. Downloading & preparing the components (git cloning...)
> 2. Building components rpms (and component dependencies to rpms)
>    a. This creates 2 yum repositories that can be used for other installs
> 
> 3. Installing component rpms (and dependents created in #2)
> 4. Starting (uses the rpms installed services, aka `service nova-compute
> start`)
> 5. And the one u want restarting!
> 
> Other actions @ 
> https://github.com/stackforge/anvil/tree/master/anvil/actions
> 
> Its also not mean for deployment (although the created yum repos can be
> used for this) but it might get u farther and suite your end-goal better.
> 
> -Josh
> 
> On 1/2/14, 5:01 AM, "Sean Dague" <sean at dague.net> wrote:
> 
> >On 01/01/2014 11:16 PM, Dean Troyer wrote:
> >> On Wed, Jan 1, 2014 at 8:58 PM, Peter Cheung <mcheung63 at hotmail.com
> >> <mailto:mcheung63 at hotmail.com>> wrote:
> >> 
> >>     1) import an qcow image to glance
> >>     2) start the instance for that image
> >>     3) init 6
> >>     4) delete that qcow image
> >>     5) rejoin the devstack
> >> 
> >>     devstack will still running, but unable to start any instance, even
> >>     instances are not from that image.
> >> 
> >> 
> >> Restarting after a reboot is really outside what DevStack is intended to
> >> be able to do.  rejoin_stack.sh is only meant to restart the screen
> >> sessions and nothing else.  Even then it isn't perfect, which is why you
> >> should run stack.sh again after rebooting.
> >
> >Right, devstack is not a generic installer, it's a development tool.
> >Surviving reboot is intentionally not a solved problem, because that's
> >not what it's for.
> >
> >	-Sean
> >
> >-- 
> >Sean Dague
> >Samsung Research America
> >sean at dague.net / sean.dague at samsung.com
> >http://dague.net
> >
> 
 		 	   		  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack/attachments/20140104/a6168148/attachment.html>


More information about the Openstack mailing list