Internal Commands

This section describes the commands which can be issued while WATCH is running. Issuing these commands will cause WATCH to switch to an alternate screen that provides more detailed information.

NOTE: You cannot edit internal WATCH commands, they must be typed correctly.

ADVISE
This command instructs WATCH to inform you of system performance statistics and sysgen parameter values that may be of interest. WATCH will advise you when any of the following events occur:
    * Any disk has a HEAVY I/O rate (more than 25 per second)
    * The HARD FAULT rate is high (above 10%)
    * Any disk has a high DISK RESPONSE TIME (above 40 ms)
    * Number of computable processes is greater than 3 (low responsiveness)
    * KERNEL mode time is high (above 25%)
    * INTERRUPT stack time is high (above 20%)
    * Cache hit percentages are low (below 70%)
    * AWSA voluntary decrementing is turned off
    * AWSA voluntary decrementing is turned on
    * Working set sizes cannot grow (WSINC is zero)
    * Page Fault Rate High/Low unreasonable (PFRATH is less than PFRATL)
    * Page Fault Rate High may be unreasonable (PFRATH is greater than 500)
    * Page Fault Rate High may be unreasonable (PFRATH is less than 10)
    * FREEGOAL is less than FREELIM
    * FREEGOAL may be too low (FREEGOAL is less than 3*FREELIM)
    * Number of free pages low (close to FREELIM)
    * AWSTIME may be too low (AWSTIME is less than QUANTUM)
    * LONGWAIT is low (less than 3 seconds)
    * LONGWAIT is high (greater than 20 seconds)

FORMAT
    ADVISE

EXAMPLE
    $ WATCH/TOP=FAULT/INTERVAL=15
    ADVISE

After starting WATCH, with or without additional parameters, enter the word ADVISE to switch you to the ADVISE screen. Watch will ADVISE you if any of the above listed conditions is true, these conditions are mentioned in the OpenVMS Guide to System Performance as potentially impacting system performance.

DISKS
The DISKS command has 2 functions. Its first one is to allow you to respecify disks from within WATCH, without having to exit and restart. Its second function is to show the disk screen. The file DISKNAMES.DAT must exist in the WATCH_EXE directory with the disknames you want to show in the disks screen.

FORMAT
DISKS/MODIFY=(dev:,...)

To modify the current list of disks being displayed on the main screen. The list you specify here will replace the existing one, be sure to list all disks you wish to see displayed.

DISKS

Switch to the DISKS screen showing more detailed statistics on all the disks listed in the WATCH_DAT:DISKNAMES.DAT file.

KEYWORDS
    dev:
    Is any diskname valid on the node that watch is running on.

EXAMPLES
    $ WATCH/DISK=(DISK1$,DISK2$)
    DISKS/MODIFY=(DISK2$,DISK3$,DISK4$)

Run WATCH showing disk statistics for DISK1$ and DISK2$, subsequently the disk display is changed to show DISK2$, DISK3$ and DISK4$.

$ WATCH
DISKS

After starting WATCH enter DISKS to change to the DISKS screen which displays usage statistics on all the disks in your DISKNAMES.DAT file.

EXIT
This command terminates WATCH. You can also use CTRL/Z to exit. The screen is cleared when you exit from the program.

FORMAT
    EXIT

KEYWORDS None

EXAMPLE
    $ WATCH
    EXIT

MONITOR
WATCH will accept any valid MONITOR command. Enter the command and WATCH will create the subprocess and go directly into MONITOR.

FORMAT
    MONITOR valid-monitor-item

KEYWORDS
    Valid-monitor-item

Refer to the OpenVMS System Management Utilities Reference Manual: M-Z (OpenVMS V6.0 documentation) or the Performance Management Guide, Monitor Utility (VMS V5.x documentation) for a description of valid monitor classes and items.

EXAMPLE
$ WATCH
Monitor processes

After starting WATCH the user entered MONITOR PROCESSES to be attached to a subprocess running the OpenVMS monitor utility to monitor the processes class.

PROCESS
This command will show the process screen. This screen is similiar to the SHOW/PROCESS/CONTINUOUS command. The screen will update itself every interval. This command can also be issued from the user screen by typing the letter 'P' and entering the process id.

FORMAT
    PROCESS/ID=process-id
    from the main WATCH display screen

    P
    process-id
    from the WATCH user screen

KEYWORDS
    process-id
    A valid OpenVMS process-id that you want to display.

EXAMPLE
    $ WATCH
    PROCESS/ID=202000A3

After running WATCH the display is switched to the process screen for process #202000A3.

    $ WATCH
    USER
    P
    000A3

After running WATCH and switching to the user screen the user enters P to switch WATCH to the process screen after entering the process number (at least the last 3 characters) when prompted.

REPAINT
This command redisplays the screen. It is used to redraw the screen after a message has been broadcast to the screen. You can also use CTRL/W to repaint the screen.

FORMAT
    REPAINT

KEYWORDS None

EXAMPLE
    $ WATCH
    REPAINT
    Repaint the screen after it was altered by a broadcast message.

SPAWN
This command creates a subprocess within the current process. SPAWN is used to temporarily leave WATCH, so that you can execute DCL commands, observe processes, and then return to WATCH. When the SPAWN command is issued a "$" appears. You can enter any normal OpenVMS command such as "SHOW QUE/FULL" and "MONITOR". When you log out of the sub_process, the WATCH screen will reappear.

FORMAT
    SPAWN

KEYWORDS None

EXAMPLE
    $ WATCH/RECORD=TODAY.WAT
    SPAWN
    DIRECTORY /SINCE *.WAT
    LOGOUT

In this example the user has started watch and specified that it record the session. To verify that recording is taking place they SPAWNed to a subprocess, did a directory and were returned to WATCH when they logged the subprocess out.

TOP
This command selects the type of jobs to be displayed in the process area. It acts in the same manner as the qualifier /TOP.

FORMAT
    TOP field_name

KEYWORDS
    "CPU"
    Show top CPU processes.

    "BIO"
    Show top Buffered Input/Output processes.

    "DIO"
    Show top Direct Input/Output processes.

    "FAULT"
    Show top page faulting processes.

    "WS"
    Show processes with the largest working sets.

EXAMPLE
    TOP FAULT

Change the process portion of the main display to show the top six page faulting process in descending order by page fault rate.

USERS
This command expands the process display to a full screen. The screen is cleared, and all the processes are displayed 20 per screen. When the screen is full, WATCH waits until you hit any key, then it continues showing processes. If the logical name WATCH$ACTIVE is defined to be anything other than WATCH$ACTIVE only processes that have used resources during the last interval are displayed - this will significantly decrease the number of processes included in the display on larger systems. Press R to return to the main screen.

FORMAT
USERS[/INCLUDE=(process_mode[,...])]

KEYWORDS

process-mode

When switching to the user screen it is possible to limit the types of processes WATCH displays by using the /INCLUDE qualifier. If the /INCLUDE qualifier is not specified the USERS screen defaults to /INCLUDE=ALL. Valid qualifier values are:

    "INTERACTIVE"
    Show only interactive processes.

    "BATCH"
    Show only batch processes.
    "NETWORK"
    Show only network processes

    "OTHER"
    Show only processes that are not interactive, batch or network processes

    "ALL"
    Show all processes on the system.

EXAMPLE
    $ WATCH
    USERS/INCLUDE=INTERACTIVE

After starting WATCH switch to the USERS screen and show information on the interactive processes on the system.