IBM FlashSystem

Hi,

Both systems should be at code level 8.6.2.x or later.
did you check the IBM Redbooks "A Practical Guide to Policy-Based Replication and Policy-Based High Availability for IBM Storage Virtualize Systems"  https://www.redbooks.ibm.com/abstracts/sg248569.html

Chapter 2.3 provides a step-by-step guide to recovery testing.

The basic principle is that it allows you to bring the target volumes online so that you can test your recovery procedures. 

This mode provides a way to test your recovery procedures without having to do a full failover of the replication.  When you do a full failover of the replication, not only is replication stopped while you do the test, but replication will have to copy all the changed data over to the target when the test completes. 

But recovery test mode allows you to return back to full replication mode faster in that it doesn't have to wait to recopy all the changed data. It's recommended to take a Snapshot at the target site prior to entering recovery test mode to ensure there's a consistent recovery copy to restore from in the event of a disaster.  

If you use Copy Services Manager (CSM) to manage the PBR relationships, CSM will automatically take a Snapshot at the recovery site for you when you enter the test mode.  It will also clean up that Snapshot automatically when you stop the recovery test mode.  

Does it means that in recovery test mode, operational data continues to replicate at DR storage, but inconsistent data is written to DR storage?
If so, where are the write data or Bitmaps stored?

Check out the following blog from Barry Whyte and Andrew Martin.  I think it does an excellent job of explaining the various test recovery options.  

https://barrywhytestorage.blog/2024/09/26/making-the-most-of-async-policy-based-replication/