RDF System Management Manual for H-Series RVUs (RDF 1.8)

su*v.fname, *.fname, and *.*, for example, are not valid.
But DB*.filename is valid because the asterisk is used as a subvolume name suffix. In this
case, changes made to all audited files and tables on all subvolumes whose name starts with DB
on the protected data volume are replicated.
Within Filenames
When used by itself as a filename, the * designates all audited files and tables on the specified
subvolume. When used as a suffix, the * designates all audited files and tables on the specified
subvolume whose names start with the supplied pattern. If the * is not used by itself to represent
all files, it must be used as a suffix. Embedded wildcard characters (such as fil*nam) are not
valid.
INCLUDE and EXCLUDE Processing
You can specify a maximum of 100 INCLUDE and EXCLUDE parameters for each volume, in
any combination. When you specify both INCLUDE and EXCLUDE clauses for a given volume,
the clauses are processed hierarchically, with the INCLUDE clauses processed first and EXCLUDE
clauses processed second.
Error Checking
Extensive checking is done when the subvolume and file names are parsed, and invalid names
cause errors. The logic of a series of INCLUDE and EXCLUDE clauses, however, is not checked.
For example, information is not regarded as an error, even though it makes no sense because
everything will be filtered out:
SET VOLUME INCLUDE MMTEST10.*
SET VOLUME EXCLUDE MMTEST10.*
Performance Ramifications
The extractor processes all INCLUDE and EXCLUDE clauses. If the extractor reads an audit
record associated with a file or table not specified in an INCLUDE clause, it discards the record.
Similarly, if the extractor reads an audit record associated with a file or table specified in an
EXCLUDE clause, it discards the record.
With volume-level replication, the extractor needs to test only the volume name to determine if
a record should be sent to the backup system. With subvolume- and file-level replication, however,
the extractor must also test the subvolume name and filename. Hence there is more work to do
with subvolume- and file-level replication. If you use large numbers of INCLUDE and EXCLUDE
clauses for each volume, the extractor might have to evaluate the subvolume name and filename
against multiple lists, which could lead to increased CPU usage by the extractor and lower
extractor performance than with simple volume-level replication. Therefore, you should be
careful about how many INCLUDE and EXCLUDE clauses you specify for each volume. The
use of wildcard characters in subvolume names and filenames can help considerably. For example,
to replicate all of the files within all subvolumes whose names begin with the letters DB, you can
do so with a single INCLUDE clause:
INCLUDE DB*.*
Summary Examples
Consider this updater configuration example, where the primary system is \PRIMARY and the
backup system is \BACKUP:
SET VOLUME CPUS 1:2
SET VOLUME IMAGEVOLUME $IMAGE
SET VOLUME PRIORITY 185
SET VOLUME PROCESS $MM01
262 Subvolume-Level and File-Level Replication