Performing a virtual switch - assure_mimix - 10.0

A virtual switch of an application group allows you to use a target node in your MIMIX production environment to test applications. If you have not already done so, it is strongly recommended that you take time to become familiar with how virtual switching affects your MIMIX environment and the testing considerations for applications. See Support for virtual switching for details.

Starting the virtual switch procedure: 

To start the virtual switch, do the following:

1. From the Work with Application Groups display, type 15 (Switch) next to the application group you want and press Enter.

   The Switch Application Group (SWTAG) display appears.

2. Specify *VIRTUAL for the Switch type prompt.

3. At the Node roles prompt, specify the node to be used for testing in the virtual switch. (No changes to the recovery domain roles will occur.) The default value *CURRENT will use the first backup node in the current node roles. This value is appropriate for typical 2-node application groups. To use a different backup node in an application groups with multiple backup nodes, specify its name.

4. Press Enter.

5. The Procedure prompt appears. Do one of the following:

   • To use the default virtual switch procedure (SWTVRT), press Enter.

   • To use a different virtual switch procedure for the application group, specify its name.

6. Press Enter.

7. You can check status of the virtual switch procedure from the Work with Procedures display. Refer to instructions in Working with procedures and steps.

   Important!

   For problem resolution, a virtual switch procedure differs from other types of procedures in these ways:

   • A virtual switch procedure is intended to stay in *MSGW state at step MXVRTRCY for the entire time that application tests are being performed on the designated test node. Do not respond to the message for this step until testing is complete and you are ready to start the recovery phase of the virtual switch. For all other steps, handle a *MSGW procedure status like that for any other procedure.

   • If a problem occurs in any step within the range performed while the application group has a replication status of *VRTSWTSTR, *VRTSWTTST or *VRTSWTRCY and the virtual switch procedure fails or is canceled, you cannot acknowledge the failed or canceled procedure. The virtual switch procedure must be resumed after you have investigated the cause of the problem for the affected step. Furthermore, you will not be able to run any other procedure to start, stop, or switch the affected application group until the virtual switch procedure is resumed and allowed to run to completion. The range of steps to which this behavior applies begins with step MXSETVRTS and applies through step MXFAILRCYV.

   • If a problem with the source node occurs that requires an unplanned switch while a virtual switch is in progress, the virtual switch procedure must be allowed to complete. This is necessary even in environments with more than one backup node in the application group’s recovery domain. If the virtual switch is in its testing phase, end the application being tested and respond to the message waiting at step MXVRTRCY. Allow the virtual switch procedure to complete before attempting to start the unplanned switch. Any recovery actions that require capturing data from the source node will fail and will have a Not Recovered status in the Virtual Switch Activity window in the Assure Unified Interface.

   Testing phase: The MIMIX environment on the designated test node is ready for application testing when the virtual switch procedure successfully processes step MXSETVRTT. As a result of this step, the Work with Application Groups display indicates that virtual switch testing is in progress with the value *VRTSWTTST in the Repl. Status column if replication processes are active. (Apply processes and access path maintenance remain ended while in the testing phase.)

   Important!While you are testing, the procedure is designed to wait for a user reply that indicates testing is complete before it processes step MXVRTRCY. You will see the status *MSGW on the Work with Procedure Status display for the duration of the testing phase of the virtual switch. Do not respond to this message until you have completed testing.

8. If you have not enabled and customized step STRUSRAPPV, you must manually start the application to be tested on the designated test node.

9. Test the application.

   While testing, it is good practice to keep a record of your test node activity that may require manual cleanup when testing is complete. This includes any setup that will need to be removed and any changed objects that are outside of the scope of replication for the participating data groups.

10. To see the objects eligible for replication that have been changed on the test node, use the Virtual Switch Activity action for the application group from the Application Groups portlet in the Assure Unified Interface.

11. For the duration of virtual switch testing, monitor the application group for potential problems from the Work with Application Groups display. Check for any indication of problems that require action and take action as needed.

12. When testing is complete, end the application on the test node.

    Note: If you enabled step ENDUSRAPPV and customized it with actions to end the application, you can skip this step. When enabled, the customized step runs after you respond to the message waiting in Step 14.

13. Perform any necessary manual clean up on the test node.

    Recovery phase: 

14. To indicate that testing is complete and you are ready for the virtual switch procedure to start its recovery phase, you must respond to the waiting message. Do the following:

    1. From the Work with Application Groups display, type 21 (Procedure status) next to the application group you want and press Enter.

    2. From the Work with Procedure Status display. type 11 (Display message) next to the virtual switch procedure (SWTVRT) and press Enter.

    3. You will see the message “Procedure name for application group name requires response. (G C).” To start the recovery phase of the virtual switch, type G and press Enter.

       Note: A response of C (Cancel) will cancel the procedure. If you cancel the procedure, you must resume the procedure and allow it to run to completion.

       This starts the recovery phase of the virtual switch and changes the replication status for the application group to indicate that virtual switch is recovering (*VRTSWTRCY on the Work with Application Groups display). Apply processes will be started and will begin processing the accumulated transactions and recovery actions.

15. You can check the progress of recovery actions from the Virtual Switch Activity window within the Assure Unified Interface.

16. When the procedure completes, check the Virtual Switch Activity window for any objects with failed recoveries. You will need to take action to address these objects.

Additional information about using the Virtual Switch Activity window to check recovery status and start manual resolution of failed recoveries is available in online help for a MIMIX instance from within the Assure Unified Interface.