Creating the RCDSs

Prior to running CICS® VR, three RCDSs must be created using the naming convention as described in the previous section. The points to consider when creating the RCDSs.

To reduce the possibility of losing the recovery information, IBM® recommends that you create each of the three RCDSs on separate volumes, disk controllers, and channels. This prevents the loss of recovery data due to a single hardware failure.
The RCDSs must be created on separate volumes, disk controllers, and channels from those of the VSAM spheres that you update and want to forward recover with CICS VR. This helps to prevent failures that corrupt your production VSAM data from affecting the RCDSs.
CICS VR contains logic to ensure that all CICS VR instances that belong to the same CICS VR XCF group use the same RCDSs. Therefore, you must create the RCDSs on storage that is accessible from all systems in a sysplex that run an instance of the CICS VR server address space that belongs to the same CICS VR XCF group. All CICS VR instances in the CICS VR XCF group need write access to the RCDSs.
In addition, you need to create separate sets of CICS VR RCDSs for each CICS VR group defined in a sysplex.
One of the RCDSs must be defined with a smaller size. Therefore, if the RCDSs are becoming full, you can be notified and take proper action prior to encountering this situation on all three RCDSs. If all three RCDSs become full, CICS VR processes terminate until proper corrective action is taken.
If this is the first time you are installing CICS VR, and you are using CICS VR VSAM batch logging or backup notification, you might want to use the following sizings as a starting point for your RCDSs.
```
DWWCON1 - CYLINDERS(200 30) 
DWWCON2 - CYLINDERS(200 30) 
DWWCON3 - CYLINDERS(150 30)
```
If you do not plan on using CICS VR VSAM batch logging or backup notification, then refer to the RCDS formula in Creating and maintaining the RCDS when not using the CICS VR server address space for RCDS sizing information.