We are just tossing ideas around with our new data center layouts. One idea that was suggested is splitting our SQL always on cluster between two data centers instead of using Zerto for replication. While we understand there is some advanced configurations that would need to be done to prevent an accidental failover – or having the 2nd node and 1st node compete for resources (witness node) – has anyone implemented it this way?
Not a direct answer to your question, but can you tell me your considerations of using versus not using ZVR to protect this server?
For example, with ZVR you gain many advantages such as hundreds+ of recovery points, within seconds of each other, and the ability to recover this from the same recovery point-in-time as other related servers if desired.
Are you looking more for high availability, plus DR against the use case of site loss? This is great, but you would be missing out on the data consistency recovery options from use cases such as corruption/blue screen/loss, isolated testing, recovery with other applications, etc. So I’m wondering what you’re looking to gain that is raising the consideration of losing other use-case scenarios.
I’m not saying either choice is bad because it does depend on your needs and goals, I’m just curious in order to gain more knowledge to help with a more detailed answer. 🙂 More info is also needed not only around your end goals (availability and/versus protection scenarios), but also site configuration, bandwidth available versus currently used, distance between sites, data rate of change, etc. These are things our team can help you with as far as replication calculators and expected RPOs as well.