Add Support for Kafka and Elasticsearch database application on Rook clusters

[parth-gr]

Is this a bug report or feature request?

• Feature Request

Supporting such database would require replica-1 topological based provisioning + Ceph's localpool functionality

Currently we have a support for replica-1 pools which will be documented soon,
So this feature request is to explore more on ceph local pools(https://docs.ceph.com/en/latest/mgr/localpool/)

What should the feature do:

Do topological provisioning on local pools

What is use case behind this feature:

Support Kafka and Elasticsearch into Rook clusters

Environment:

[parth-gr]

@travisn localpool looks like, distributing the data across the cluster,

This can be useful for deployments where it is desirable to distribute some data locally and other data globally across the cluster

But for replica1 we need individual pools specific for a failure domain, not think if the idea will be suited,
Any comments what do you think on this?

[travisn]

From the doc:

For example, by default, it will create a pool for each distinct rack in the cluster. This can be useful for deployments where it is desirable to distribute some data locally

I'm still not clear if it's useful for our scenario, but it does indicate that pools can be created within failure domains.