Re: [Xen-devel] [PATCH v4] tools: set migration constraints from cmdline

[Ian Campbell <Ian.Campbell@xxxxxxxxxx>]

On Mon, 2013-02-04 at 13:55 +0000, Olaf Hering wrote:
> On Mon, Feb 04, Ian Campbell wrote:
> 
> > > Instead the migration should be aborted and the guest should continue to
> > > run on the old host.
> > 
> > That's what I understood from reading the code, can we make the docs say
> > it too ;-)
> 
> -A --abort-if-busy maybe?

Could work. Not sure if everything needs to have a short option, but I
suppose there's no harm.

> > > > >  =item B<remus> [I<OPTIONS>] I<domain-id> I<host>
> > > > > diff -r fd711ebdce9a -r 785c8f34e1f8 tools/libxc/xc_domain_save.c
> > > > > --- a/tools/libxc/xc_domain_save.c
> > > > > +++ b/tools/libxc/xc_domain_save.c
> > > > 
> > > > The changes to this file only seem to implement the -N and not the other
> > > > two?
> > > 
> > > xc_domain_save already has max_iters and max_factor as arguments.
> > 
> > Ah, I didn't appreciate that.
> > 
> > BTW did you want to add a way to override DEF_MIN_REMAINING?
> 
> I did consider it, but havent looked closer how to do it.

Might be best to do while you are changing the API anyway?

It just occurred to me, instead of adding lots of individual arguments
perhaps packing them into a (e.g.) libxl_save_properties and adding a
pointer would be a nicer and more extensible (in the future) interface?

> > I'm not really sure what the usecase is for save -c. I'd be a bit
> > inclined to leave it until someone wants it I think.
> 
> save -c looks like a snapshot feature, but does not seem to take storage
> into accound. So yes, I will leave it alone.

Ack.



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