|
|
|
|
|
|
|
|
|
|
xen-devel
Re: LAST_CHECKPOINT and save/restore/migrate compatibility (was Re: [Xen
On Sat, Apr 2, 2011 at 1:13 AM, Ian Campbell <Ian.Campbell@xxxxxxxxxxxxx> wrote:
On Sat, 2011-04-02 at 04:57 +0100, Shriram Rajagopalan wrote:
> On 2011-04-01, at 6:58 AM, Ian Campbell < Ian.Campbell@xxxxxxxxxxxxx> wrote:
> > Is it the case that Remus only cares about checkpointing between like
> > versions of the toolstack?
> >
> Can you please elaborate this statement ?
For standard suspend/resume or migration we support migrating from
version N to version N+1 (but not vice versa), to support upgrades.
In the case of Remus though are we interested in supporting a rolling
checkpoint from a version N system to a version N+1 fallback system? Or
is Remus only interested in supporting checkpoints between systems
running the same version of Xen?
N->N+1 should work, for Remus even now, without the LAST_CHECKPOINT patch. But as you said, failback from N+1 -> N wont work.
I agree with George's suggestion on MORE_CHECKPOINTS, for backwards compatibility wrt live migration. But for HA, it doest make much sense if a user is able to do HA only one way and cannot failback. This is not a upgrade scenario. So, that would require some
exception to be thrown when a version incompatibility is detected. IMO, its better to let the user handle this limitation than letting him/her do the N->N+1 HA and then finding out that they cannot failback.
shriram
Bear in mind that if you did support N->N+1 checkpoints you wouldn't be
able to migrate the VM back after a failover...
FWIW I think George got to the bottom of the specific issue he was
seeing and that the LAST_CHECKPOINT thing was a red-herring, although
flipping the protocol to use a MORE_CHECKPOINTS schema perhaps makes
sense anyway.
Ian.
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel
|
|
|
|
|