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

Table 3 Software Requirements (continued)
RequirementSoftware
The RDF/IMP, IMPX, and ZLT products use Expand software
to connect the primary system to the backup system.
Communications
On the primary and backup systems, the installed release
version update (RVU) of the operating system must be
supported.
Operating System
On both the primary and backup systems, the installed
RVU of the TMF subsystem must be compatible with the
installed RVU of the operating system.
TMF Subsystem
On both the primary and backup systems, the installed
RVU of the NonStop SQL product must be compatible with
the installed RVU of the operating system.
NonStop SQL Products
Configuring TMF for RDF Operations on the Primary System
TMF attempts to purge old audit trail files each time it rolls over to a new one. The purge is
performed only if the audit trail file is not pinned on behalf of RDF.
RDF automatically pins audit trail files. The only ways TMF can purge an old audit trail file that is
still required by RDF are:
• If you issue an RDFCOM UNPINAUDIT command while RDF is not running.
• If you stop TMF and restart it without restarting RDF. TMF does not retain pinning on behalf
of RDF when TMF is stopped and then restarted. If you must stop and restart TMF, be sure to
restart RDF before you restart your applications. This causes RDF to re-pin the audit trail files
it needs, and thereby prevents TMF from purging the files before RDF has finished processing
them.
If you issue an UNPINAUDIT command while audit dumping is disabled and TMF purges an audit
trail that has not yet been processed by RDF, you will have to reinitialize RDF and resynchronize
the databases. If you have configured TMF for audit dumping, however, you will not need to
reinitialize RDF or resynchronize the databases (the extractor will wait until the needed audit trail
is restored and then resumes).
AUDITTRAIL BUFFER
After you have configured your TMF audit trails, for each audit trail disk you should configure the
AUDITTRAILBUFFER ON and configure it with a reasonable value. You do this with the SCF utility
program. By default, AUDITTRAILBUFFER has a value of 128 megabytes, and this may well be a
reasonable value for you. By configuring AUDITTRAILBUFFER to at least 128 megabytes, you allow
the extractor to read the audit trail files from disk cache rather than physically from disk. Also, by
setting the buffer to a high value, you allow the extractor reads to continue to go to cache instead
of to disk even when the extractor has fallen behind. Note, if the extractor should fall way behind,
for example the communications line to the backup system fails, and if you have insufficient cache,
then the extractor's reads will go to disk until it catches up to what is in cache. The ability to read
from cache is clearly a performance gain for optimal extractor-to-receiver throughput. Please note
that altering the value of AUDTITRAILBUFFER can be done offline or online, but if you do it online
your new value will not take effect until you take the disk down and then bring it back up.
TMF Configuration With Dump Process on the Primary System
When you configure TMF with audit dump on, that subsystem dumps an audit trail file to tape or
disk before purging the audit trail file. This approach is strongly recommended on the primary
system.
Audit trail files are pinned by the RDF extractor and TMF cannot purge pinned files until the extractor
has finished processing them. TMF will keep these files pinned on behalf of the RDF extractor even
Preparing Software and Database Files for RDF Operations 55