AWS AZ fail over?

[stv.r...@googlemail.com]

Hi, 

I understand that replication/mirroring is available across AWS AZs, but from the documentation I'm not 100% clear how this works in practice - hoping you can help.  

My interpretation of the docs is that redundancy is via HA, but is this still the case if we have 3 nodes running in zone A, and that AZ fails over (but no existing nodes/replication in zone B), will the storage be available to new zone B nodes - considering there's no nodes from zone A to sync from?

Thanks for your help

Steve

[Travis Nielsen]

To support zone failure in the same cluster, you will need to have running OSD(s) in each of the zones. Then you create a pool with “failureDomain: zone”, such as found here: https://github.com/rook/rook/blob/master/cluster/examples/kubernetes/ceph/pool.yaml#L13-L14

As long as the nodes have the zone labels (as should be the case by default in aws) and the pool created with the zone failure domain, one replica would be placed by Ceph automatically in each of the three zones. 

For further questions, we also invite you to join the Rook slack.

Travis

Rook Maintainer

--
You received this message because you are subscribed to the Google Groups "rook-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to rook-dev+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/rook-dev/18c5bf89-9b09-4055-b750-ddd31457e0cdo%40googlegroups.com.