Hi all,
I have been researching WIN 2016 features and new AGs enhancements lately. Need your guidance on the following requirement and proposed hypothesis for my virtual lab purpose. I'd like to ensure that my interpretation of AG concepts is accurate. Appreciate further comments and alternate approach if my hypothesis is not applicable.
OS: Win2016; SQL; 2016 EE
Requirement:
Let's assume I have 4 nodes(N1,N2,N3,N4), 2 nodes in Location A (site 1) and other 2 in location B (site 2). App1 is accessing (A,B,C) databases in site1 for all their operations and App2 is accessing (X,Y,Z) databases in site 2. In my case, all sites act as primary to their business apps. All write operations should be independent to each site whereas reads could be load-balanced.
Following are my proposals based on my understanding about FCI/AGs and Strech Clusters
-- Using FCI and AGs
1) Create single WSFC between 4 nodes
2) Create and configure FCI-1 on nodes N1 and N2
3) Create and configures FCI-2 on nodes N3 and N4
4) Create AG-1 and AG listener from site1 to site 2 on databases (A,B,C) - readable secondaries (async or sync)
5) Similarly AG-2 and listener on X,Y,Z databases from site2 to site1
Hypothesis: It has HA between N1 and N2 within site1 and at the same time I have secondary replicas pointing to site2. I could also setup Readable Routing list (FCI-2) to make all reads load- balanced across two FCIs. Vice-versa applies to another set of databases from site2 to site1.
Harsha