incubator-deltacloud-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Chris Lalancette <>
Subject Re: Question about instance states
Date Mon, 25 Oct 2010 17:59:10 GMT
On 10/22/10 - 03:57:15PM, Ed Bradford wrote:
> I noted Chris Lalancette's email on Oct 7 saying there are only three
> states defined.
> However, many of the drivers still reference "shutting_down". I have a

Yes, I did not cleanup the state machines.  So the API can no longer return
a state of SHUTTING_DOWN, but these things are still listed in the files.
Note, however, that these are orthogonal issues, really; it's possible that the
backend state machine could go through a "SHUTTING_DOWN" state, but we just
never expose that to the API user.

> suggestion. Would it be
> possible to include in the source code of each file the date on which it
> was
> last modified? That would be very helpful finding why something is missing
> or present.

This is trivial with git:

git blame <filename>

> [I could be wrong on this. I downloaded the source code Oct 22, 2010 from
> git clone git:// core]
> Also, on the "states".
> What is the current thinking? Specifically, :pending is so vague that it
> will always
> beg for more information. Pending what? If I am writing software to talk to
> Deltacloud
> and my instance is "pending" what can I extract from that information? I'm
> thinking
> that :booting, :shutting, :suspending [or :pausing] and maybe
> even :snapshotting
> give much more information to the user than :pending.

Yes, I agree that pending is vague.  But I'm not sure that we have a good
way to tell some of these states; EC2, for instance, does not give us any
additional information except PENDING, and then RUNNING, once it is up.

That being said, I'm not at all opposed to adding new states.  But it needs
to be done in a consistent way across the drivers, otherwise it is impossible
to use the API in a cloud-independent way.

Chris Lalancette

View raw message