SPI Common Extensions Manual

ZCMK Definitions
SPI Common Extensions Manual—427508-001
B-7
Selecting a Log File
A value of FALSE, RESET, OFF, or NO disables debugging.
LOG filename or
LOGFILE filename
specifies the name of the log file. For important considerations related to log file
selection, see Selecting a Log File on page B-7.
If you omit the parameter or the filename argument, the subsystem does not
generate logging output. Emergency output is sent to the home terminal.
If you specify an asterisk (*) in place of the filename argument, logging output is
sent to the standard output file.
RECEIVEDEPTH integer
specifies the number of entries reserved for incoming messages on the file
$RECEIVE. The value of the argument integer must be in the range 0 through
2000, and should be based on anticipated data traffic. (The subsystem augments
this value to account for system messages and SPI requests.)
If you omit this parameter, the subsystem-defined default is used (see the
subsystem-specific management programming manual). If the subsystem does not
provide a default value, the maximum value (2000) is used.
SWAPVOL volname
specifies the name of the disk on the local node that is used for swapping
extended memory. The demands for space placed on that disk are discussed with
the DATAPAGES argument. If omitted, the volume specified in the SWAP process
option is used. If that volume is also omitted, the disk from which the program was
loaded is used.
Selecting a Log File
A ZCMK-based process uses the log file specified in the LOG or LOGFILE startup
parameter, or the file specified in the most recent ALTER command. Regardless of
how you specify it, consider the following points when selecting a log file.
Do not use a terminal as the logging file unless special circumstances require it.
Logging uses waited I/O operations, and terminals are frequently unavailable because
of manual input operations. This can adversely affect subsystem process performance.
A terminal might be an appropriate log file in testing and debugging environments but
usually not in production environments.
Using a printer as the log file is subject to the same concerns. Printers can become
not-ready, and the resulting interruption of logging activity can severely impact
performance.
Using a spooling process as the log file lets you examine the resulting files online using
the Peruse program or printed after an ALTER command switches logging to another
file.