Can replay trace be used to sync data between SQL instances?

Category: sql server distributed


aurorahokie on Thu, 11 May 2017 22:42:45


We are using SS2K12 behind an on prem MSCRM install. I have AG's configured for our databases on identical hardware etc.. Unfortunately our CRM releases take a lot of time and we risk data loss due to CRM needing to be offline to client requests during these deployments. This is causing the dev team to work nights and weekends, which is what I'm trying to help prevent. 

Here's the plan...

1) Break AGs
1) Bring a secondary CRM server online (as well as a node from each of the other farms involved)
3) Start replay trace on primary node still taking client traffic
4) Run release against the "B" side and allow dev to test
5) Once release is complete we point customers to the upgraded "B" side (B side doesn't have any data/transactions for the duration of the release and testing at this point)
6) run replay against b side to sync the data.
7) Move all farms into appropriate pool and pointed to the B side then set up AG's again

Has anyone tried this? Am I crazy for trying?

We are NOT highly transactional, so I'm assuming 100MB trace data at most (but need to verify).