[CRIU] Remote lazy-restore design discussion

[Mike Rapoport]

On Tue, Apr 05, 2016 at 07:06:30PM +0300, Pavel Emelyanov wrote:
> On 04/05/2016 11:52 AM, Adrian Reber wrote:
> > On Mon, Apr 04, 2016 at 04:06:50PM +0300, Pavel Emelyanov wrote:
> >> On 03/31/2016 05:25 PM, Adrian Reber wrote:
> 
> Yup. And I also like your initial proposal to work in "pull mode" for the first
> 5 seconds. I think it will let us keep the network free from unneeded pages while
> the program is re-warming after lazy restore.

The exact policy for things like "when to start background page transfer",
"how many pages should be fetched together with the faulting one" is a
subject for another long discussion :)
 
> I've sent more detailed view of how this all should look like in another e-mail
> in this thread. I think that this 5 seconds pull-only period would fit there.
> 
> -- Pavel

--
Sincerely yours,
Mike.