Open Stack

I have an OS setup with multiple compute availability zones, but only a
single cinder storage AZ currently.  I have a few questions on this:

1) Are there any rules/optimizations/defaults implemented in the OpenStack
components about availability zone symmetry between compute and block
storage services?  I can attach volumes to any of my compute nodes
regardless of region, so I presume there isn't a strict restriction, but
are there pitfalls to attaching instances to volumes with different AZ
names?

2) What's considered the standard for handling availability zones for
different service types? Should a 1:1 compute AZ to block storage AZ be the
goal?

3) Is there any way to define rules governing which cinder availability
zones compute nodes can attach volumes from based on the compute
availability zone?


-- 
Andrew Mann
DivvyCloud Inc.
www.divvycloud.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack/attachments/20140626/73fce361/attachment.html>