[occi-wg] Question about Compute instance action

[Ralf Nyren]

Sounds good!
/Ralf

On Mon, 12 Sep 2011 21:16:47 +0200, Edmonds, AndrewX  
<andrewx.edmonds at intel.com> wrote:

> Agreed. Perhaps as a note I can append "termination of a resource  
> instance is performed using HTTP DELETE (see section x.y.z)"?
>
> -----Original Message-----
> From: Ralf Nyren [mailto:ralf at nyren.net]
> Sent: Monday, September 12, 2011 4:28 PM
> To: Edmonds, AndrewX
> Cc: Andy Edmonds; Kim Rohner; occi-wg at ogf.org
> Subject: RE: [occi-wg] Question about Compute instance action
>
>
> Yes, agreed on the non-intuitive part :)
>
>
>
> Is it provider specific what happens if you http-delete a running Compute
>
> instance? It could be equivalent to the stop-terminate case...
>
>
>
> Using Actions to Create/Delete Resource instances seems to push to usage
>
> of actions a bit to far IMO.
>
>
>
> /Ralf
>
>
>
> On Mon, 12 Sep 2011 14:13:51 +0000, "Edmonds, AndrewX"
>
> <andrewx.edmonds at intel.com> wrote:
>
>> A method - the transition from inactive to terminal state would be
>
>> 'terminate'. The action of 'stop' and http-delete is certainly the only
>
> way
>
>> to do it currently but as Kim's question shows, this is perhaps not
>
>> particularly intuitive.
>
>>
>
>> -----Original Message-----
>
>> From: occi-wg-bounces at ogf.org [mailto:occi-wg-bounces at ogf.org] On Behalf
>
>> Of Ralf Nyren
>
>> Sent: Monday, September 12, 2011 4:05 PM
>
>> To: Andy Edmonds
>
>> Cc: Kim Rohner; occi-wg at ogf.org
>
>> Subject: Re: [occi-wg] Question about Compute instance action
>
>>
>
>>
>
>> Would the "stop-terminate" state be equivalent to a stop Action followed
>
>> by a removal of the Compute instance?
>
>>
>
>> The state diagram in [1] page 6 has shows the "stop" action changing the
>
>> state from Active to Inactive. You mean to add another state Andy?
>
>>
>
>> regards, Ralf
>
>>
>
>> On Mon, 12 Sep 2011 15:25:49 +0200, Andy Edmonds <andy at edmonds.be>
>
> wrote:
>
>>> Hi Kim,
>
>>> The required functionality is in the specification as an Action against
>
>> an
>
>>> instance specified by its URL. Please see Section 3.1 (table 3, 2nd
>
> row)
>
>> of
>
>>> the Infrastructure specification [1] and also the Core specification
>
> [2]
>
>>> for
>
>>> more details on Actions. 'stop' will result in termination. In [3]
>
> there
>
>>> are
>
>>> examples of such Actions (e.g. pg 11). On review of the noted sections
>
>>> there
>
>>> is no explicit specification of a stop-terminate state change as you
>
>> note.
>
>>> It's a good catch and the Infrastructure specification will be updated
>
>> to
>
>>> reflect this, should no one else here have an objection to the change?
>
>>>
>
>>> Thanks!
>
>>>
>
>>> Andy
>
>>> andy.edmonds.be
>
>>>
>
>>> [1] http://ogf.org/documents/GFD.184.pdf
>
>>> [2] http://ogf.org/documents/GFD.183.pdf
>
>>> [3] http://ogf.org/documents/GFD.185.pdf
>
>>>
>
>>> On Fri, Sep 9, 2011 at 16:43, Kim Rohner <KimRohner at web.de> wrote:
>
>>>
>
>>>> Dear OCCI-Workgroup,
>
>>>>
>
>>>> i have a question concerning the OCCI Compute instance actions.
>
>>>>
>
>>>> Comparing to the Amazon EC2-API,
>
>>>> OCCI obviously delivers no action to terminate an Instance.
>
> Terminating
>
>>>> in the Amazon-API means, that the instance is shut down and won´t be
>
>>>> visible after one hour.
>
>>>> The Action you listed in the GFD-P-R.184 on page 5 gives only the
>
>>>> possibility to hibernate or suspend.
>
>>>>
>
>>>> Is that the same or am I just confused :) ?
>
>>>>
>
>>>> Thanks for clearing that up for me !!!
>
>>>>
>
>>>> Feel free to answer in German :)
>
>>>>
>
>>>> Keep up the good work,
>
>>>>
>
>>>> Kind regards,
>
>>>> Kim Rohner
>
>>>> _______________________________________________
>
>>>> occi-wg mailing list
>
>>>> occi-wg at ogf.org
>
>>>> http://www.ogf.org/mailman/listinfo/occi-wg
>
>>>>
>
>> _______________________________________________
>
>> occi-wg mailing list
>
>> occi-wg at ogf.org
>
>> http://www.ogf.org/mailman/listinfo/occi-wg