nova team mailing list archive

[Paul Voccio <paul.voccio@xxxxxxxxxxxxx>]

They should be consistent between them as the scheduler would need to
determine which host/cluster a new request should go to and where to route a
subsequent request for that vm/host.

Shouldn¹t the layer be pluggable if I wanted a different datastore? Wouldn¹t
putting it all in the scheduler¹s memory and managing it there make that
more difficult?


On 8/4/10 1:37 PM, "Michael Gundlach" <michael.gundlach@xxxxxxxxxxxxx>
wrote:

> Eric,
> 
> On Wed, Aug 4, 2010 at 2:20 PM, Eric Day <eday@xxxxxxxxxxxx> wrote:
>>> >    In your new arch diagram (
>>> http://oddments.org/tmp/architecture_new.png)
>>> >    isn't the host/guest cache the same thing as the datastore? The
>>> scheduler
>> 
>> The host/guest cache is the in-memory table with data it needs
>> about hosts, clusters, and guests to answer some queries and route
> 
> If the scheduler service is a cluster of replicas, will they all duplicate the
> entire table in memory?  Do the tables need to be consistent between them (do
> hosts have to push updates to every scheduler worker?)  Would it be simpler to
> pull that out into a memcached that they share?
> 
> Michael
>

Attachment: smime.p7s
Description: S/MIME cryptographic signature