Monitor conf/texis.ini Section

The settings in the [Monitor] section of conf/texis.ini change general properties of the monitor processes.

Run Level

Default: 1
Sets the run level of the Texis Monitor. Its value is an integer whose bits indicate the following: bit 0 is whether to run as overall Texis Monitor, bit 1 is whether to exit if the default database is removed. Internal/unsupported use. Added in version 3.01.981800000 Feb 9 2001. See also this setting in the [Scheduler] section.

Log File

Default: %LOGDIR%/monitor.log in version 8 and later (%INSTALLDIR%/texis/monitor.log in version 7 and earlier)
The file that monitors should log to. This can be overridden with the -o command-line option.

Lock File

Default: %RUNDIR%/dbmonitor.lck in version 8 and later (%INSTALLDIR%/texis/.dbmonlck in version 7 and earlier)
On non-Windows platforms, this is the name of the lock file that database monitors check as an indicator of whether to exit when monitor -k is issued. This file and its directory should be writable by the Texis user. Added in version 3.01.985300000 Mar 22 2001.

Log Native IO

Default: ignored (treated as 0) in version 8.01.1711127229 20240322 and later, else 0
On Windows platforms, if nonzero, use native system calls for log file I/O. Generally enabled only at request of tech support. Deprecated (ignored, treated as 0) in version 8.01.1711127229 20240322.

Log Reopen

Default: [Texis] Reopen value in version 8.01.1711127229 20240322 and later, else 0
If nonzero, re-open the log file for every message, instead of keeping it open between messages. This allows system log rotators to rotate monitor log files without restarting the monitor, allowing disk space to be recovered even if the monitor is stuck. This is a legacy setting; if modification needed, generally [Texis] Reopen (here) is set instead (to control all log files, not just monitor's).

Log Thread Id

Default: permanently on in version 8 and later; no in version 7 and earlier
Whether to log the thread ID (if not main) for every message. This setting is ignored in version 8 and later, as the (non-main) thread ID is then always logged.

Pid File

Default: unset
If set, the file that the Texis Monitor writes its process ID to. Generally for debug use; monitor -k determines PID from the license shared-memory segment.

Trace Dns

Default: 0
Trace DNS calls according to given integer's bit flags. Same format as the Vortex <urlcp tracedns> setting. For debugging/tech support use.

Trace Socket

Default: 0
Trace socket calls according to given integer level. Same format as the Vortex <urlcp traceskt> setting. For debugging/tech support use.

Trace Pipe

Default: 0
Trace pipe calls according to given integer level. Same format as the Vortex -tracepipe command-line option. For debugging/tech support use.

Trace Fcgi

Default: 0
Trace FastCGI calls according to given integer level. For debugging/tech support use.

Timestamp

Default: 0
If nonzero this will write a time stamp to the log file every Timestamp seconds. This can be used to make sure the monitor is still running.

Refresh

Default: 1
If you are running a monitor at a terminal, and there is an interactive display (e.g. ltest), this setting sets the default refresh interval for the display. This can be modified while the program is running by hitting a numeric key.

Keyboard Read

Default: 0.1
This is related to Refresh, and controls how often the keyboard is checked for input.

Semaphore

Default: 10
This value (in seconds) controls how often the semaphore is checked to make sure that it has not become stuck, which would cause the database to be unusable.

Semaphore Timeout

Default: 10
This value (in seconds) controls how long the semaphore should be ungettable before it is considered stuck.

Removal

Default: 30 (60 in version 6 and earlier)
How often the database should be checked for removal. The default is 30 seconds; prior to Texis version 7 it was 60 (1 minute). If you frequently create and delete databases this will prevent the database monitor process from running too long.

DB Quiet

Default: 120
Sets the time in seconds that the database monitor should keep watching the database if there are no accesses to the database. Once this idle time has elapsed with no accesses, the database monitor will exit. This value plus Removal should be larger than any of the settings that periodically access the database, e.g [Chkind] Refresh or [Monitor] DB Cleanup Interval, so that the latter do not artificially cause the database to appear too-recently-used when DB Quiet looks for idleness.

Upgrade SYSTEM Tables

Default: 10 (0 in version 7 and earlier)
Interval in seconds (0 for never) to attempt to upgrade older existing databases' system tables to the latest schema, e.g. whether to add the PARAMS column to the SYSINDEX table, and upgrade the SYSUSERS table to varchar to enable longer users and more secure passwords (in version 8 and later). Upgrading is recommended, as these columns are needed for some newer features. Added in version 3.01.992053000 20010608.

Statistics

Default: 3600
How often the database statistics should be updated, in seconds.

Stats Block

Default: 0
If nonzero, a blocking connect() call will be used by the statistics monitor client. This is normally 0 (off), except under Linux 1.x kernels (no longer supported) where it defaults to on to avoid a bug in the kernel.

Mem Limit

Default: -1
Virtual-memory limit for monitors: if a monitor process exceeds this limit it will exit. -1 for no limit. Can have MB etc. suffix, e.g. 100 for one hundred megabytes. Used for debugging.

Fork

Default: 0
Whether to attempt to fork() the monitor program to start it when possible (Unix), for certain instances. Bit 0 controls forking for monitor -s (internal statistics gathering), bit 1 controls monitor -C.

Max Scheduler Fails

Default: 5
Maximum number of failed attempts to start Vortex <schedule> jobs before exiting. Debugging use. Note that a failed or non-zero-exiting script is not generally considered a failed attempt at starting Vortex jobs.

Verbose

Default: 0x5
Integer whose bit flags control some log messages:

• 0x0001: Database monitor start/stop

• 0x0002: Semaphore removal

• 0x0004: Windows service control

• 0x0008: Check stats start/stop

• 0x0010: More check stats messages

• 0x0020: Startup errors

• 0x0n00: Same as n occurences of monitor -v flag

These flags are subject to potential change in a future release.

Use Ddic Mutex

Default: yes
Boolean: Whether to use a mutex to protect internal DDIC usage. Turning this off can potentially cause monitor problems, and is generally recommended only at the request of tech support. Added in version 5.01.1239305000 20090409.

DB Cleanup Interval

Default: 180 (60 in version 6 and earlier)
Integer number of seconds between database cleanups, which look for deleted and temporary indexes and tables and try to remove them (if no longer in use). This cleanup also happens automatically as needed - e.g. before index creation - but cleaning up periodically may save some disk space sooner. Added in version 6.00.1338325000 20120529. The value should be less than [Monitor] DB Quiet (see discussion there).

DB Cleanup Verbose

Default: 0
Integer whose bit flags control some log messages about database cleanup housekeeping (e.g. removal of unneeded temporary or deleted indexes and tables) when conducted by the Database Monitor. A bit-wise OR of the following values:

• 0x01: Report successful removal of temporary/deleted indexes/tables.

• 0x02: Report failed removal of such indexes/tables.

• 0x04: Report on in-use checks of temporary indexes/tables.

Note that these cleanup actions may also be handled by any Texis process that accesses the database; see also the dbcleanupverbose SQL property. Added in version 6.00.1339712000 20120614.

License Flush

Default: allpossible
Level at which to flush license.key license file after writing; one of:

• none No flushing

• data Flush data to disk (error if unsupported)

• all Flush data and metadata to disk (error if unsupported)

• allpossible Flush data and/or metadata, if possible (some platforms unsupported)

Using a value other than none can help prevent problems on reboot due to a missing or corrupt license, if the machine is shutdown improperly (e.g. power failure). Added in version 7.06.1506612435 20170928.