Re: [Xen-devel] [OSSTEST PATCH 23/27] make-flight: Provide xen-unstable-smoke branch

[Ian Campbell <ian.campbell@xxxxxxxxxx>]

On Wed, 2015-09-16 at 15:51 +0100, Ian Jackson wrote:
> Ian Campbell writes ("Re: [OSSTEST PATCH 23/27] make-flight: Provide xen
> -unstable-smoke branch"):
> > On Wed, 2015-09-16 at 14:35 +0100, Ian Jackson wrote:
> > > This contains a very limited set of jobs:
> > >     build-amd64
> > >     build-armhf
> > >     test-amd64-amd64-libvirt
> > >     test-amd64-amd64-xl-qemuu-debianhvm-i386
> > >     test-armhf-armhf-xl
> > > 
> > > The debianhvm job exists only in this flight, and is generated by
> > > having branch_debianhvm_arch return i386 instead of amd64.
> > 
> > any particular reason for this?
> 
> I wanted to test at least one 32-bit guest.  I have added a note about
> this to the commit message.

OK.

I wonder if we should also add this to the regular flights? Seems odd to
smoke test it for unstable but not for e.g. real 4.6 flights.

> > > Deployment note: This requires images/debian-7.2.0-i386-CD-1.iso
> > > which I have already downloaded to the Cambridge instance using
> > > jigdo.
> > 
> > But not the main colo?
> 
> Not yet.  I will do that before I start trying to run this there.  I
> think I ought to do such a --real run, by hand, when the series seems
> about ready to go into pretest.
> 
> > > +case "$branch" in
> > > +xen-unstable-smoke)
> > > + global_runvars+=" hostalloc_maxbonus_variation~=0 "
> > > + global_runvars+=" hostalloc_bonus_previousfail~=0 "
> > 
> > Why are these synthetic? 
> 
> I wrote in the commit message that introduced the ~ notation in
> cs-job-create:
> 
>     This will be useful if we want to set hostalloc_* runvars.  Normally
>     we would want to set those only on flights generated by make-flight.
>     
>     Flights generated by cs-bisection-step or cs-adjust-flight ought not
>     to copy them, because cs-bisection-step makes its own arrangements
> for
>     host specification, as should the caller of cs-adjust-flight (perhaps
>     via the blessing system).
>     
>     Using `synth' for this is arguably slightly wrong but it does the
>     right thing in all existing cases.  The alternative would be a schema
>     change.
> 
> Perhaps I should move some of this text to this commit message.

Yes, I'd forgotten all about it by now, which I think it is when it is
relevant.

> 
> Ian.

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel