Stopping TMF

Before a system fallback, stop TMF twice with no database activity between the two stops:

TMF 1> STOP TMF
			<stop tmf output>
TMF 2> START TMF
			<start tmf output>
TMF 3> STOP TMF
			<stop tmf output>

These steps are necessary to avoid volume recovery when TMF is started after the system fallback. TMF recovery of SQL/MX database objects will access the RCBs of the affected objects. You must shut down TMF because the RCBs of the affected objects are not immediately compatible after a change from 3.5 to an earlier release.

NOTE:

Do not fall back from 3.5 immediately after a system crash. In case of a system crash, you must restart the system with the version that was running at the time of the crash, start TMF, and then shut down TMF twice before falling back.