HP NonStop RDF, Руководство по управлению

Страница: 194 / 478

Usage Guidelines
With the ADD command, all configuration parameter settings that you previously supplied in
SET commands for the particular process or other object are applied from the RDF memory table
to the RDF configuration file. Any parameter settings that you did not supply are set to their
default values.
Each volume on the primary system protected by RDF requires a corresponding updater process
on the backup system. You must issue one ADD VOLUME command for every primary system
volume to be protected by RDF.
Although more than one primary volume can be protected by a single disk volume on the backup
system, HP strongly recommends that you configure your environment so that each updater
process writes to its own volume (that is, that there is a one-to-one mapping of primary volumes
to backup volumes).
If you add an updater process after you stop RDF, the context record of the new updater process
is set to that of the receiver’s context for particular image trail of the updater. So, after the next
START RDF command, the newly added updater process begins reading from the current image
file at the same place that the receiver process begins writing.
CAUTION:
After RDF is configured and running, do not add an updater process unless a STOP
TMF command has shut down RDF; otherwise, you cannot be sure that the data on the newly
added backup system volume is synchronized with the data on the corresponding primary
system volume.
The master image trail identified by the RDFVOLUME parameter in the SET RECEIVER command
is reserved for use by the MAT receiver process. All updaters must be configured to secondary
image trails. Each image trail is stored on a separate volume on the backup system. To add a
secondary image trail, you specify the disk volume intended for its use through the ADD
IMAGETRAIL command. When you configure your individual updater processes with the SET
VOLUME command, you assign each of these processes to a different image trail, reducing the
number of updaters contending for a specific trail.
Each secondary image trail contains only that audit data required by the associated updaters.
Image files in secondary image trails have the same extent sizes as image files on the volume
specified by RDFVOLUME. Because the extent sizes for secondary image trails are obtained from
the receiver’s configuration record, you do not specify them through any RDFCOM commands.
With RDF/IMP, IMPX, or ZLT, if your RDF configuration is not particularly small (say, 30 volumes
or more) and your performance requirements are not high, you can assign one image trail per
updater process, and you can place that image trail on the updater’s UPDATEVOLUME. If you
have the need for high performance, however, even on a small system, it might be best to have
a dedicated volume for your image trails. For an RDF environment involving very high replication
rates, you should not configure more than 3 or 4 updaters to an image trail if the replication rate
of these updaters is high.
You cannot add secondary image trails until you have configured the receiver, because the extent
sizes for these trails depend on those specified for the receiver’s master image files.
Examples
To configure the extractor process named $EXT to run in CPUs 0 and 1 at the default priority of
165 with the default RTD warning threshold of 60 seconds, enter the following commands:
]SET EXTRACTOR PROCESS $EXT
]SET EXTRACTOR CPUS 0:1
]ADD EXTRACTOR
To define $SYSTEM.RDFIMP as the location of the RDF software, enter the following commands:
]SET RDF SOFTWARELOC $SYSTEM.RDFIMP
]ADD RDF
194 Entering RDFCOM Commands