Micro Focus recommends you use the following best practice guidelines when configuring and using
Performance and Availability Cluster (PAC)s:
- Ensure that traffic to your PAC is managed by a load balancer. This will enable you to perform any future product upgrades
in a way that minimizes PAC downtime and loss of data.
- TSQs and TDQs should be stored in their own SOR and not in the PSOR. This is for performance reasons and to facilitate the
product upgrade process.
- Each enterprise server instance requires its own resource definition file (dfhdrdat). Only a subset of resources are stored
in the
PAC Scale-Out Repository (PSOR). An enterprise server instance still require resources that are not available from the
PSOR. The same dfhdrdat file should be used in every enterprise server instance in the
PAC.
- The same SIT and startup list should be used for every
enterprise server region in the PAC.
Note: SIT information should be the same, except for the SYSID if it is used for CICS inter-system communication (ISC).
The APPLID should also be unique as this becomes the nodes netname, and the SYSID which is used internally to uniquely identify
the connected systems within the other
enterprise server regions.
- Use the same program and file paths across a PAC.
- Only use the
caspac command to modify entries in the
Scale-Out Repository (SOR). Removing entries manually might lead to unexpected behavior.
- Prior to cold starting a PAC, ensure that you have run the
caspac command to initialize the PAC.
- Do not use Redis to permanently store all applications.
Micro Focus recommends that you only store applications that are to be modified whilst a PAC is active and those changes need to be deployed
throughout the PAC.