I'm a little nervous replaying a captured trace/workload for performance analysis...

I'm using an isolated service account to run the SQL Services, including the Distributed Replay controller & clients....Is that all that is needed to ensure replayed transactions won't jump across to the prod database?

IE, is it possible that procedures or CLR objects could reach out and connect to another database, even though no linked servers are defined?



Since the target SQL Server could have a stored procedure using a hard-coded SQL Server account and password to connect to some other server, I'd be unwilling to be 100% certain nothing could touch production unless I had a truly separate network setup to run the Distributed Replay batches against.

One could go to great lengths to try to ensure nothing hits production, but what if you miss something?

I'd setup a bunch of VMs for the controller, replay machines, and the target SQL Server(s), and configure them on their own virtual network, just to be sure.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.