RDF System Management Manual for J-series and H-series RVUs (RDF Update 13)

When the extractor receives notice of the operation, it notes where it is in the audit trail and shuts
down, and the updaters shut down as soon as they have reached the equivalent location. This is
identical to the DRAIN command.
Next, RDF automatically executes the REVERSE trigger that you have configured. You would want
this trigger to accomplish the same types of actions as the TAKEOVER trigger discussed further
below in #9 in the section “How to Plan for the Fastest Movement of Business Operations to Your
Backup System After Takeover” (page 133). Unlike with the TAKEOVER trigger, you would want to
include in the REVERSE trigger the RDFCOM commands to configure a new RDF subsystem to run
from your former backup system to your former primary system and then start that subsystem.
CAUTION: If you fail to stop the application associated with your RDF subsystem in step 1, the
probability is high that you will corrupt your backup database. Then, if the REVERSE trigger moves
application processing to the backup system and update activity is replicated by RDF back to the
original primary system, you may find it very difficult to bring the two databases back into
synchronization without loss of committed data.
For further information about the STOP RDF, REVERSE command and the REVERSE trigger, see the
corresponding sections in “Entering RDFCOM Commands” (page 175).
Reciprocal Configurations and Switchover
In a reciprocal RDF configuration, two systems act both as a primary and as the backup to the
other.
With reciprocal configurations it is imperative that you make sure the file-sets being replicated by
the two RDF subsystems are absolutely independent of each other, and this can only be done in
one of two ways:
1. The volumes protected by RDF Subsystem #1 are completely different from the volumes protected
by RDF Subsystem #2. For example, RDF Subsystem #1 protects volumes $DATA1-$DATA10,
and RDF Subsystem #2 protects volumes $DATA20-$DATA30.
2. You use INCLUDE/EXCLUDE lists to ensure complete separation. For example, RDF Subsystem
#1 INCLUDES SUBVOLA.* on volumes $DATA1-10, and RDF Subsystem #2 INCLUDES
SUBVOLB.* on volumes $DATA1-10.
Please note that in each of these configurations, there is no possible overlap of data being replicated
by the two RDF subsystems in a reciprocal configuration. See Reciprocal and Chain Replication
in Chapter 1 for a discussion of the potential problem that can occur if you do not use one of the
two methods described.
The steps for performing a planned switchover from \A to \B in such a configuration are:
Carrying Out a Planned Switchover 127