ASAP 3.2 Server Manual

HP NonStop ASAP Server Manual Page 131 of 381
find the file or process. ASAP does provide counts of objects when viewing the
wildcard domains at aggregate levels.
Wild-card names can resolve into many thousands of objects. That number of records
can put pressure on disk resources when storing historical data, so ASAP
automatically adds an aggregate-only domain to limit the output from wild-card name
specifiers after you issue a MONITOR command that contains a wild-card name. You
must manually delete the aggregate-only domain for individual records on each object
to be written to the historical database. For example if you enter MONITOR PROCESS
$A*, then ASAP issues a MONITOR PROCESS $A*\## command to add an
aggregate-only domain. For more information on aggregate and aggregate-only
domains, see the subsections on aggregation later in this section.
Wild-card domains that are resolved at each interval are stored in a memory pool in
the File and Process SGPs. The pool is allocated from extended memory and is
controlled by KMSF. Excessive numbers of objects can deplete the pool space. The
MONITOR FILE | PROCESS, LIST, DETAIL command can be used to view pool
usage. The MBYTES parameter to the ASAP File and Process SGP processes is used
to control the amount of space within the pool. For more information, see the
MONITOR Command, and the FILE and PROCESS options of the SET Command.
To set an objective on the wild-card domain, you must include the wild-card specifier in
the name. For example, to set an objective on the Full attribute for process $ABC that
was resolved from the $A* wild-card specifier, you enter RANK PROCESS $A*\$ABC,
FULL < 90. You can also set a global objective at the wildcard level, for example
RANK PROCESS $A*, BUSY < 95.
Monitoring Processes from an Object File
ASAP supports monitoring of all processes running from an object file. The object file
can be a Guardian or OSS object filename. When object filenames are specified,
ASAP resolves the name at each interval to determine the processes that are
executing from that object file. ASAP computes and creates a record for each process
found. To avoid name conflicts with other forms of name specifiers ASAP appends the
process name to the end of the name specified as a new level of the ASAP domain
name. For example, the command MONITOR PROCESS $DATA.OBJECT.FILE would
resolve into a domain named $DATA.OBJECT.FILE\$ABC for process $ABC.
Because the set of objects resolved from an object filename can vary at each interval,
ASAP does not issue an alert if it finds a member of the set is no longer running or
present. This contrasts the way ASAP works if you directly monitor a process using the
process name. In that case ASAP does issue an alert when it cannot find the process.
ASAP does provide counts of objects when viewing the domains at aggregate levels.
Object filenames can resolve into many objects. That number of records can put
pressure on disk resources when storing historical data, so ASAP automatically adds
an aggregate-only domain to limit the output from object-filename specifiers after you