On Tue, Dec 16, 2014 at 04:50:18PM -0000, Ryan Harper wrote:
> I spent quite a bit of time to see if there was something to do. The
> current state of pxe control in QEMU isn't ideal. You can disable
> loading of the option rom which prevents any nic of that *type* from pxe
> booting. However, there is no control over a per-nic basis.
@rharper -- didn't you also say something about a bridge timeout?
>
> The changes included switch the first nic to use virtio (instead of
> rtl8139, unrelated but a better, faster choice), and then switch the
> second nic to be e1000, and then include the <rom file=''> directive
> which disables pxe booting on all e1000 nics. The VM now will pxe boot
> only from the virtio nic. If the bridge is slow or maas is busy, then
> the VM may not successfully pxe boot. This may or may not be more
> desirable from an Orange Box perspective.
>
It should put us into a 'fast-fail' sitation, and remove a
false-positive (i.e., the node looks fine in MAAS, but is not reachable
via name or IP listed in maas).
If we are using juju -- the problem gets detected as a timeout
eventually (also not ideal). After switching to this, it would get
detected earlier as a 409, and a 'failed deployment' in the MAAS GUI.
I think all-in-all it's a good change for the orange box as it will
allow us to more precisely detect when the problem occurrs.
On Tue, Dec 16, 2014 at 04:50:18PM -0000, Ryan Harper wrote:
> I spent quite a bit of time to see if there was something to do. The
> current state of pxe control in QEMU isn't ideal. You can disable
> loading of the option rom which prevents any nic of that *type* from pxe
> booting. However, there is no control over a per-nic basis.
@rharper -- didn't you also say something about a bridge timeout?
>
> The changes included switch the first nic to use virtio (instead of
> rtl8139, unrelated but a better, faster choice), and then switch the
> second nic to be e1000, and then include the <rom file=''> directive
> which disables pxe booting on all e1000 nics. The VM now will pxe boot
> only from the virtio nic. If the bridge is slow or maas is busy, then
> the VM may not successfully pxe boot. This may or may not be more
> desirable from an Orange Box perspective.
>
It should put us into a 'fast-fail' sitation, and remove a
false-positive (i.e., the node looks fine in MAAS, but is not reachable
via name or IP listed in maas).
If we are using juju -- the problem gets detected as a timeout
eventually (also not ideal). After switching to this, it would get
detected earlier as a 409, and a 'failed deployment' in the MAAS GUI.
I think all-in-all it's a good change for the orange box as it will
allow us to more precisely detect when the problem occurrs.
Thanks for this nice write-up, btw.
--
David Britton <email address hidden>