Not sure if this is a tech mistake I've made or a lack of understanding:
- Architecture = 2 x 2 node clusters, with 1 x AG each & 1 x testDB in AG1.
- Aim = to use distributed availability group to port testDB from clusterA to clusterB.
Steps:
- Create DAG on clusterA listener to listener.
- Join to DAG on clusterB. Grant create any database.
- Looks good. testDB appears on clusterB / node 1. Failover is fine - happy with it all.
My problem is that stepping back to pre-failover, the database is never seeded to clusterB node2. The database is not created there through AG2. It does appear in the AG under availability databases but the join fails as the database is not created. There are no errors to suggest an issue, which makes me think this is probably how it is supposed to work.
Thanks.