Switch processing checks for the existence of replicated jobs on the current source system as a condition that ends switch processing. Any replicated jobs on the current source system are from a previous switch and are a condition that will end a planned switch when defaults are used.
In application group environments, the switch procedures also wait for you to either submit or remove any remaining previously replicated jobs on the new source system following the switch and set an indicator if any replicated jobs still exist for the affected data groups.
Replicated jobs are addressed by the following enabled steps within procedures for switching application groups. You can also use any of the optional, disabled steps listed to have the switch procedure address any remaining jobs following a switch.
|
Step |
Step Description |
Notes |
|---|---|---|
|
Enabled Steps |
||
|
MXCHKRJB |
Checks for replicated jobs. This step checks to see if there are remaining replicated jobs on the current source node of the data group. This step fails if there are any remaining replicated jobs. |
Data group level step, runs on primary node before the actual switch operation occurs. Enabled in PRECHECK and, SWTPLAN procedures. |
|
MXWAITRJB |
Wait for replicated jobs to be submitted/removed. This step waits for all replicated jobs on the new source node to be manually submitted or removed. |
Data group level step, runs on new primary node to provide the opportunity to address any replicated jobs on the new primary node after the switch operation occurs. Enabled in SWTPLAN and SWTUNPLAN procedures |
|
MXSETRJBI |
Set replicated jobs indicator. This step sets an indicator if there are remaining replicated jobs on the new source node for the data group. The data group cannot be switched again until all remaining replicated jobs are submitted or removed. |
Data group level step, runs on new primary node after the switch operation occurs. This step cannot be disabled or removed. Enabled in SWTPLAN and SWTUNPLAN procedures |
|
Optional, Disabled Steps |
||
|
MXSBMRJB |
Submits all replicated jobs. This step submits all replicated jobs on the new source node, allowing the jobs to run. |
Optional, shipped disabled in SWTPLAN and SWTUNPLAN procedures. Note: User action is required to enable these steps. See Enabling or disabling a step.
When enabled, each of these data group level steps run on the new primary, have a before action of *WAIT, and have an error action of *MSGW. |
| MXSBMRJBN |
Submits all new (not started) replicated jobs. This step submits all replicated jobs that had not yet been started on their original source node to run on the new source node. |
|
|
MXRMVRJB |
Removes all replicated jobs. This step removes all replicated jobs from the new source node to prevent the jobs from being submitted to run. |
|
For more information about procedures and steps, see procedures and step programs.
In data group-only environments, the Conditions that end switch (ENDSWT parameter) on the Switch Data Group (SWTDG) command supports the *REPJOB value, which will cause the switch process to end if any replicated jobs exist on the source system. In a planned switch, this is also included with the value *DFT.