Checking group switch readiness using RTVSWCHRDY
The old joke used New Yorkers on tourists who ask “How do I get to Radio City Music Hall?” – to which the response is always “practice, practice, practice”, also applies to high availability. To ensure you have a valid backup system that you can switch to in case of a disaster or for planned maintenance, you need to practice your role switches again and again and again!
To determine the switch readiness of a group, if you are on any node at iCluster 7.1 you can look at the status monitor. Check to see if the group is active, has little to no latency, has no suspended objects and no out of sync objects. If all of these look good, the group is ready for a switch. If you want to do this through a batch process or CL program, you can use the RTVSWCHRDY command.
RTVSWCHRDY prompts for the group name ICGROUP, and then CL variables of type *CHAR LEN(1) for RDYSTATUS and STATUSINFO. These variables will contain information on the status of your group that will be passed back to your program.
RDYSTATUS will contain one of the following vales:
0 – The group is ready for switchover
1 – The group is not valid for a role switch – see the STATUSINFO value for additional information
2 – The group has suspended objects
3 – The group has out of sync objects
4 – The latency has exceeded the user thresholds
5 – There are open commitment control cycles on the backup node
STATUSINFO will contain one of the following values:
0 – No additional information is available
A – Not all group information could be retrieved. Look for previously listed messages in the joblog.
B – The group uses local loopback replication
C – The group or one of its object specifiers has a target library that is not set to *PRIMARY.
D – The group is a refresh only group.
Finally the command prompts for whether out of sync objects should be ignored (ALWOOS set to *YES) or whether suspended objects should be ignored (ALWSUS set to *YES). You may want to set these values to *YES if you know for example that you have obsolete or extra objects on the backup node that will not be part of the role switch test. Stay tuned for more on switch readiness in the next post!