For HA configuration i would recommend primary server in one data center/fault domain and secondary on secondary data center /different fault domain. This will ensure high availability in case of any network or power failure. It is always a best practice to isolate Dev/Test from production servers. If dev/test is added to PI production collective then in case of any issues with PI version (upgrade/change) in test/dev server then collective will be out of sync with production servers HA.
If HA is configured with Production and Dev/Test then Interfaces writing data to PI collective has to write to 3 different nodes using buffer subsystem and latency+interface server configuration to be considered. Secondly, If your primary server is subjected to high configuration changes then it will result high bandwidth usage. PI Client will connect to one server at any point in time.
High availability collective : Information related to failures and high level considerations.
Please let me know in case of any queries.
Great answer, Lal!
You will run into many issues trying to use a member of a PI Collective as a test/dev server. Lal mentioned a few reasons already, such as the fact that all members of a PI Collective must be running the same version of the Data Archive. Additionally, you introducing risk to your production data availability and storage.
I would highly encourage you to configure two servers in an HA Collective, and use a 3rd, standalone server as your dev/test system. We have some recommendations and best practices listed here for working with a test system: KB01163 - PI System in a Test Environment