In EC2, you can stop a server, and after 60min, EC2 itself will delete it. You cannot restart a stopped instance on EC2.
For Rackspace, you could implement the "shutdown" action, which also implies deletion, perhaps.
There's the instance_states FSM that determines which actions are applicable to which instances in which states. That's up to the specific driver to construct. That's where you can say "an instance can do X Y and Z, but not W"
wrt Storage Volumes, yeah, that's still TBD and modelled only after EC2 at the moment. We need to address mountable volumes soon across the board.
Sounds like Rackspace's persistent backups would be similar to rebundling an EC2 image?
-Bob
On Sep 22, 2009, at 5:32 AM, Michael Neale wrote:
Hi All. In implementing the Rackspace driver, I noticed that I can't stop a server (only delete it) - so is there some way we should allow the driver to throw/return something that says that operation is not supported at all for a given fabric? (until we find a way) ? (or have a list returned of what is allowed, to the Framework?).
Also - I noticed that in the driver code there is storage volumes and snapshots (once again, not directly supported, although there are a few options) - this is not mentioned in the page on building a driver in the doco (just a missing bit? or is it optional/non important anyway ? In rackspace, all instances are persistent with backups, slightly different to perhaps EC2 and other fabrics).
Thoughts?
-- Michael D Neale home: www.michaelneale.net blog: michaelneale.blogspot.com _______________________________________________ deltacloud-devel mailing list deltacloud-devel@lists.fedorahosted.org https://fedorahosted.org/mailman/listinfo/deltacloud-devel