RDF System Management Manual for J-series and H-series RVUs (RDF 1.10)

6. Install the new RDF software in a different volume.subvolume from that housing the current
version of RDF that is running. For example, if you are upgrading to T0346ABS, you might
specify $system.rdfabs.
7. Run $system.rdfabs.RDFCOM and initialize a new RDF configuration, using:
• The suffix parameter (such as suffix "a")
• The INITTIME parameter, using the timestamp calculated in the preceding example
(11AUG2008 05:24).
Initialize RDF, backupsystem \RDF06, suffix a, inittime 11AUG2008 05:24
8. If you do not already have a copy of the configuration script used for the current version of
RDF, you can get it by starting the RDFCOM for that RDF subsystem and using the INFO *,
OBEYFORM command.
9. Use the same script to configure your new RDF subsystem, but you will need to change the
following:
a. Set SOFTWARELOC to $system.rdfaav.
b. Set the extractor name(s) to a different name(s).
c. Set the monitor name to a different name.
d. Set the receiver name(s) to different name(s).
e. Set the purger name to a different name.
10. Now start your new RDF subsystem:
] run $system.rdfaav.rdfcom rdf04a
] start RDF, update off
You now have parallel sets of extractors shipping audit to parallel sets of receivers for the two
operating RDF subsystems, although each subsystem has its own control subvolumes and its
own imagetrail subvolumes.
11. When the extractor(s) for RDF04A have caught up, do the following:
a. Issue a STOP RDF command for the previous RDF subsystem.
b. Issue an UNPINAUDIT command for the previous subsystem.
c. Issue a START UPDATE command for the RDF04A subsystem. Wait until all updaters have
caught up.
d. Purge the previous control subvolumes on the primary and backup, as well as the
imagetrails for the previous subsystem.
You have now installed and started new RDF software without jeopardizing disaster-recovery
protection by having to stop RDF.
Disaster Points
If the primary system fails between steps 1 and 10, you perform the takeover operation using your
previous RDF subsystem. If the primary system fails at or after step 11, you perform the takeover
operation using the new subsystem (RDF04A).
Considerations
This method does not work with long-running transactions. You must not have any long-running
transactions in the system when you start Step 1, above. If you have long-running transactions,
you must stop them and wait until they clear the TMF subsystem before you start Step 1.
If you are running with RDF process lockstep, you should change the RDF gateway startup script
to reference the new extractor name before executing Step 11. Then stop the gateway manually.
This action will restart the gateway, and the gateway will access the new extractor.
For RDF network environments, you should subtract an additional 15 minutes from the timestamp
you calculated in Step 4.
76 Installing and Configuring RDF