With the separation of compute and storage I was expecting something similar to Snowflake where storage and the database as a whole would be separated from the cluster. I’m able to detach and attach databases in the clusters on the managed service but how can I do it across clusters? Ideally I’d want to detach a database in one cluster than attach it to another cluster.
I see in the docs that I’ll need a remote database name but I’m not able to find in the docs or anywhere that specifies this name. Is this currently possible?
You can’t detach a database yourself and attach it to a different DB right now. Our SREs can help you do this in some situations but it’s not supported in all cases - you can’t move a database from one cell to another.
You can work around it with backup and restore (but with more data copying). That you can do by yourself without help from our SRE team.
You can also grow your cluster in the portal, and the data will automatically rebalance. That can be done online.
There’s some very cool stuff our roadmap for this year related to elasticity and sharing of unlimited storage databases; I don’t want to give more details until it ships.
When thinking about detaching a DB from one cluster and reattaching to another, what goal do you want to achieve?
The main goal we have is to remove noisy neighbors when they become too loud and/or migrate them to a more intensive cpu cluster to handle their workload.