We have recently added AAG to our production servers, for DR purposes. We have approximately 60 FCI clustered instances (SQL 2017) hosting ~5000 client databases. This is all hosted in our own datacenter. The FCI's are pairs with an on-duty and off-duty node, failing over at the beginning of and mid-month. These are the primaries in the AGs. The newly added secondaries are single server Iaas VMs hosted in Azure, 1 per FCI. The AGs are set to manual failover, asynchronous commit, and will never be failed over to except in a disaster recovery scenario. The question is, how do we apply SQL updates in this scenario? I believe the steps should be:
- Suspend data movement for all DBs in the AG
- Apply update to the AG secondary and restart
- Resume data movement
- Apply update to the off duty FCI node of the Primary and restart
- After the next scheduled failover of the FCI, apply update to the now off-duty FCI node and restart
Hopefully my question is clear. I look forward to hearing from whomever has some input.
Thanks!