RDF System Management Manual

Preparing the RDF Environment
HP NonStop RDF System Management Manual—524388-003
2-7
Preparing Databases for RDF Protection
TMF Configuration Without Dump Process
Long ago, the RDF product required that you configure TMF with a dump process that
dumps to tape. RDF no longer imposes this requirement on either the primary or
backup systems for the following reasons:
1. On the primary system, the RDF extractor explicitly pins the audit trail it is currently
processing, thereby preventing TMF from purging it. This explicit pinning remains
in effect even if the extractor process fails or RDF is shut down.
If you must unpin one or more audit trail files, you can do so by issuing an
UNPINAUDIT command. Later, when RDF is restarted, you can restore the
necessary audit trail files from tape.
2. TMF includes the functional capability of audit overflow volumes. You should
always configure them with at least one overflow audit volume.
3. If you are required to do a takeover, it is recommended that you take online dumps
of the backup database before restarting the applications that will use it.
Preparing Databases for RDF Protection
When preparing databases on the primary system for RDF protection, you must
consider the following system aspects:
•
Copies of files for the backup database
•
DSM catalog and file code 900 replication
•
Copies of NonStop SQL/MP and NonStop SQL/MX views on the backup systems
•
Placement of partitioned Enscribe files and NonStop SQL/MP and NonStop
SQL/MX tables
Audited Backup Database Files
The backup system must have copies of all files that RDF protects. For a successful
takeover of business operations in the event of a primary system failure, the backup
system should also have copies of all the files needed by the primary system
applications (including alternate key files and index files, for example). For each
audited data file that resides on the primary protected volume, a corresponding audited
file must exist on a volume configured for an updater process on the backup system.
The volume name on the backup can differ from that on the primary, but the subvolume
names must be identical on both systems. For example, if volume $B on the backup
Caution. Although RDF no longer requires you to configure TMF with a dump process that
dumps to tape, you should nevertheless configure TMF for dumping to tape if you want to
achieve full TMF protection for your primary database. In addition, if the RDF extractor is
running behind and you stop the TMF and RDF subsystems before RDF has caught up to the
TMF shutdown point, when you subsequently restart TMF, the TMP might roll over the files
before the RDF extractor can process them.