Frame Migration #51
Comments
Is this for expanding the cluster? |
Yes -Todd
|
@tgruben I was trying to implement this in-process but it gets ugly really fast. The server essential needs two configurations to try to manage and then it has to cut over. I'm looking at doing a much simpler solution where we spin up a separate cluster and do a restore into the new cluster from the old one. Nodes in the new cluster, btw, can live on the same machine but just have different ports. Although since you're on AWS it probably makes the most sense to create a cluster with new machines, restore to that cluster, switch over the load balancer, and then remove the old machines. The benefits of this approach are:
We'll need to be able to have the input stream write to the old and new clusters though. @tgruben @travisturner how does that sound? |
Closed via #62. |
Remove link to developer agreement since we don't have a CLA for this project
Add "having" support to GroupBy() queries
Need some process to move a fragment to a new node. Worse case an offline process, ideally completely online without noticeable service interruption
The text was updated successfully, but these errors were encountered: