I've updated the specification for the Distributed Scheduler blueprint
to include details of the proposed implementation. Up until now it hasn't been
possible to provide any detail, since this depended on the implementation of
the changes for multi-cluster/zones. Now that Sandy has made
Good stuff. Couple of questions:a) Looks like one of the major drivers is the rules based scheduling - mutual exclusion, temporal and spatial rules, and so forth. Is there an existing way to capture these placement pragmas ? Or should we have a light rules framework ? b) Saw your note "I am looki
On Mar 10, 2011, at 11:35 AM,
wrote:
> a) Looks like one of the major drivers is the rules based scheduling -
> mutual exclusion, temporal and spatial rules, and so forth. Is there an
> existing way to capture these placement pragmas ? Or should we have a light
> rules framework ?
I have implemented a (single-node) constraint-based / rules-based scheduler
that attempts to find a "good" solution to potentially conflicting rules. I
used it to implement eday's "openstack:location=machine1.rack1.room1.dfw"
type pragma that we discussed in the past. I think this could helpful f
The OpenStack community is proud to announce the opening of registration for
our April 26-29, 2011 Design Summit conference in Santa Clara, CA at the Hyatt
Regency. Confirmed speakers for the first 2 days of the event include NASA,
Wikimedia, eBay, Intel, Citrix, Cloudscaling, GigaSpaces, KT, G
This is in regards to lp732924.
There currently isn't any simple way to know which resources are in
which project. Most resources return the project via some attribute:
* Instances: ownerId
* Addresses: instanceId
* Security Groups: ownerId
Volumes return an owner via status, but it's the user,
I have a feeling that authn/authz will try to make this more
consistent, but that won't be landing until after Cactus.
-jay
On Thu, Mar 10, 2011 at 6:47 PM, Ryan Lane wrote:
> This is in regards to lp732924.
>
> There currently isn't any simple way to know which resources are in
> which project.
The three possible solutions that I can see at the moment are:
1. Change the status return value (which seems risky, depending on how
this is used)
2. Put the projectid into the EC2 tags (which would mean getting shot by
the 'metadata mafia')
3. Support filtering (which helps Ryan,
8 matches
Mail list logo
