Re: [Xen-devel] libxl: error: libxl.c:2150:libxl_set_memory_target new target 0 for dom0 is below the minimum threshold

[Ian Campbell <Ian.Campbell@xxxxxxxxxx>]

On Wed, 2011-11-23 at 11:46 +0000, Teck Choon Giam wrote:
> On Wed, Nov 23, 2011 at 7:34 PM, Ian Campbell <Ian.Campbell@xxxxxxxxxx> wrote:
> > On Wed, 2011-11-23 at 11:30 +0000, Teck Choon Giam wrote:
> >> On Wed, Nov 23, 2011 at 7:07 PM, Ian Campbell <Ian.Campbell@xxxxxxxxxx> 
> >> wrote:
> >> > On Wed, 2011-11-23 at 10:56 +0000, Stefano Stabellini wrote:
> >> >> On Tue, 22 Nov 2011, Teck Choon Giam wrote:
> >> >> > >> >> That means default value changed if it is unset for autoballoon?
> >> >> > >> >> Original one is commented with #autoballoon = 1 in 
> >> >> > >> >> /etc/xen/xl.conf so
> >> >> > >> >> it is unset I guess so the default value for autoballoon for 
> >> >> > >> >> c/s 23110
> >> >> > >> >> is autoballoon=0 where c/s 23190 is autoballoon=1? ÃÆ?Ãâà Just 
> >> >> > >> >> some
> >> >> > >> >> guessing... ...
> >> >> > >
> >> >> > > Autoballoon and dom0_mem are incompatible.
> >> >> > > I don't think there are any relevant differences between 23110 and 
> >> >> > > 23190
> >> >> > > on xen-unstable, maybe you used to start dom0 without dom0_mem 
> >> >> > > before?
> >> >> >
> >> >> > Nope... all my servers will always have those dom0_mem set.  It is
> >> >> > from xen-4.1-testing not xen-unstable for the two changeset.
> >> >>
> >> >> I still cannot see anything in that range. However as I said before, it
> >> >> is expected that with dom0_mem set autoballoon needs to be disabled.
> >> >
> >> > We still haven't seen the full node IDs of the changesets which I asked
> >> > for so it is not obvious that we are actually looking at the same range
> >> > of changesets...
> >>
> >> Sorry, I am actually not sure what ID you are asking :(
> >>
> >> Changesets details as below:
> >>
> >> changeset:   23110:4d5c76248de3
> >
> > The "23110" bit here is only locally relevant to the repo you are
> > looking at so you cannot reliably identify a changeset to someone else
> > using that number. Only the "4d5c76248de3" is a global identifier and
> > can therefore be exchanged.
> 
> Understood :)
> 
> >
> >> user:        Roger Pau Monne <roger.pau@xxxxxxxxxxxxx>
> >> date:        Sat Jul 23 09:01:25 2011 +0100
> >> summary:     xend: remove PCI device listing from NetBSD, since it's Linux
> >>
> >> changeset:   23190:5a00ccfc6391
> >> user:        Stefano Stabellini <stefano.stabellini@xxxxxxxxxxxxx>
> >> date:        Fri Nov 18 13:38:05 2011 +0000
> >> summary:     x86: re-inject emulated level pirqs in PV on HVM guests
> >> if still asserted
> >
> >
> > Unfortunately I still don't see anything relevant between these
> > changesets so I've no idea why this seems to be a new issue. Are you
> > _positive_ that you didn't change anything else about your system
> > configuration on upgrade?
> 
> I don't change settings once it is deployed and in production use for
> xen servers.  In fact, all will be identical just the amount of memory
> to allocate to dom0 are different depending on the server total amount
> of RAM.

The message is (I think) triggered based on the %age of memory allocated
to dom0 so if different machines are different sizes and therefore have
different thresholds that might explain why it is only seen sometimes.

> > That sounds like a bug, there's no need to wait for an rc release to
> > report a bug.
> 
> I think I reported once before... let see... here:
> http://old-list-archives.xen.org/archives/html/xen-devel/2011-04/msg00972.html

I'm sorry this appears to have slipped under the radar at the time. I
think it would be worth reposting in a new thread.

Ian.

> 
> Thanks.
> 
> Kindest regards,
> Giam Teck Choon



_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel