[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [MirageOS-devel] Lifecycle



Sure, in principle -- I was wondering about a concrete example,
particularly given what I understand intended / common uses of
unikernels to be, ie., in micro-service / serverless environments.

On 19 July 2016 at 20:09, Nik Sultana <ns441@xxxxxxxxx> wrote:
> I think if we're asking to restart a service, then it might be expedient to
> avoid certain cleanups or deallocation at the stopping phase, since the same
> allocations or state will (optimistically speaking) be recreated soon at the
> starting phase. This could save wasted effort saved at the stopping phase.
> Presumably there would need some kind of timer to ensure that the full
> cleanup and deallocation does take place eventually, if the start fails.
> Cold vs warm vs tepid restart?
>
>
> On 2016-07-19 20:01, Richard Mortier wrote:
>>
>> On 14 July 2016 at 18:16, Daniel Bünzli <daniel.buenzli@xxxxxxxxxxxx>
>> wrote:
>>>
>>> Le jeudi, 14 juillet 2016 à 18:10, Daniel Bünzli a écrit :
>>>>
>>>> Since you have to deal with that idea anyway I'd say it's better not to
>>>> support a restart event.
>>>
>>>
>>> It could however be added as a *hint* to the stop event payload.
>>
>>
>> Concretely, what would be done with that hint though?



-- 
Richard Mortier
richard.mortier@xxxxxxxxxxxx

_______________________________________________
MirageOS-devel mailing list
MirageOS-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/cgi-bin/mailman/listinfo/mirageos-devel

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.