openstack team mailing list archive

[zebra z zebra <zebra.x0r@xxxxxxxxx>]

Hi.

I'm experimenting with some odd fundamental ideas about storage currently
for my deployment.

Earlier on this list, we were discussing a way to "pin" or segregate certain
instances to certain node types. There appears to be a quite neat way to do
this using a filter-scheduler configuration and a prefix on the nodes
themselves.

If we take it a step further, I'd like to know if there is a way we can run
instances that use local-to-the-node disk storage for boot and shared
storage for the vm stores at the same time. I know it's an odd question that
might be missing some logic somewhere.

What I'm trying to achieve is an availability requirement in different
modalities and different scalability requirements with the same
infrastructure, in terms of the storage used.

Thanks.

-z