I’ve got a 4 node server cluster configured where 3 servers are local and 1 server is remote. Workloads are ran both locally and remotely. The end goal is to be able to have remote workloads stay running during a local outage.
I tested an outage by severing the connection between the 2 sites, and while the local workloads stay running the remote workloads stay running for a short while, and then presumably are caught in GC when the node goes onto
No cluster leader state.
Bringing back the connection restores the consensus across the 4 nodes, and remote workloads are brought back up.
What would I need to change in my deployment to allow the remote workloads to survive the local outage?