maas-devel team mailing list archive

[Mark Shuttleworth <mark@xxxxxxxxxx>]

On 18/09/14 16:01, Gavin Panella wrote:
> On 15 September 2014 22:34, Graham Binns wrote:
> ...
>> So as it stands I'm leaning towards option #3. Questions, thoughts and
>> comments are welcome.
> During the 1.7 Feature Review call today, Kiko noted this discussion and
> brought us to a consensus:
>
> * Don't queue anything.
>
> * If there's a power command already in progress, reject the new
>   command, and return an error to the caller.
>
> This is not ideal, and that is well understood. However, it is simple
> and achievable in time for release, and provides unsurprising behaviour.
>
> Is this going to cause problems for anyone, stakeholders especially?

This is a first step to a real queue. It forces the OTHER guy to wait
and queue, but it's effectively feedback that a queue exists (without
then queueing your request).

I.e. this step:

   $ maas power-off node6
   FAILED because we are rebooting node6 for john.

Is a first step to:

  $ maas power-off node6
  DEFERRED because we are rebooting node6 for john. Then
  will power-off node6 for you.

Mark

Attachment: signature.asc
Description: OpenPGP digital signature