Probably doing something wrong, but here’s my scenario: we have an application that supports both SQL and Oracle.
Deploying it requires a long series of steps, and a couple of them involve setting the proper connection string.
One of our servers is dedicated to testing this application and it belongs to an environment called QA.
Initially, I was toying with the idea of having 2 distinct roles for this target (WebServer-Sql and WebServer-Oracle), and also had a look at this post: http://octopusdeploy.com/blog/variable-specificity-and-complexity.
But I guess the 2 roles would be useless: apart from being non-deterministic, when I deploy a new release to QA it would only hit that target once, whereas I really need to end up with 2 instances, each configured with the relevant connection string.
Am I correct in thinking that the only way to achieve my goal is to split the server into 2 targets (i.e. configure 2 tentacles on it)?
Thanks in advance,