Conference csc32::consolemanager

    
    Q1:
    
    When/how are new child console controller processes be created.
    Currently I see only 1 Ctrl process. In de documentation on 1-4 it
    looks like every process has its own child process.
    

    A1:

    For every 16 systems in your database, you will see a new child
    process started

    Q2:
    
    The documentation states to setup a poort on a DECserver as a
    printerport. It would be much appreciated to give examples in the
    documentation of what those settings can be.
    - For TELNET ports there are a couple of ways to setup such a port.
    - For LAT can the VCS settings be used ?

    A2:

    Which Terminal server would you like to se the example for?
    DECserver 200, 300, 90, 90TL, 90M or perhaps a CISCO or XYPLEX,
    also, which version of the terminal server software on each
    of these devices would you like the example for?
    
    Do you now see why we dont supply an example!

    Q3:
    
    Security:
    Can the old LAT groupcode still be used ?
    How about security on TELNET ports ?
    Sure security is an issue in worldwide connected consoles. If CM is
    down and my backupsystem is not up fast enough the line should not be
    up for grabs by any hacker.

    A3:

    Yes you can still use the LAT groupcode, just remember to set it up
    before you start PCM

    I suggest that you lobby the Terminal Server engineering folks for
    whatever security mechanism you want to see. I dont know of a way to
    achieve what you want today.
    
Cheers,
Phil

    A1:
    OK
    
    A2:
    Yes I already did see the problem but still, some guidance in the
    documentation for say the most sold 2/3 DIGITAL DECservers till that time
    and/or 1 or 2 3rd party servers you tested would be appreciated.
    
    A3:
    Thought of that, it is (in DIGITAL's case) not a problem of PCM
    engineering I know.
    
    
    Q4:
    Where can I find the logging of which dispatch was triggered by which
    event from which node. Checked
    SYS$MANAGER:CONSOLE$C3_SYSTEM_MSG.LOG
    SYS$LOGIN:C3_'username'_MSG.LOG
    CONSOLE$LOGFILES:CONSOLEMANAGER.LOG
    
    Regards, Marcel.

  Marcel,

  In answer to Q4, you cant! we do not keep this
  kind of audit trail, you could in theory turn
  on debugging for ENS which would give you this
  (plus a lot of other useless stuff), but be
  careful, the debugging stuff will be changing
  for the next release and it will burn A LOT of
  disk space.

  Why do you require this information? Maybe it
  should be submitted as a phase 0 request to Rae
  Kung our product manager.

Cheers,
Phil

    Phil,
    
    you keep nice track off all kinds of counters. In a production
    environment it is required to see as much as possible as simple as
    possible.
    
    How many times which events on which systems were trickered during the
    period of 17:00-08:30 hours (or any time), and did the action routines
    finished OK or did it bumm out so check logfile xyz.
    We want to minimize the time which have to be spent in finding and
    solving problems. Also management wants to know how many times things
    go wrong. Maybe it is necessary to modify procedures (technical or
    organization) and solve the problem at the origin. If you can produce
    numbers you can show it to management and maybe get time to solve it.
    Now you have to rely on people looking to screens and get colorblind of
    all the events passing by.
    
    Like in POLYCENTER System Watchdog when the logfile is enabled you can
    find which action routines were fired. Though this is not possible via
    SNS commands you can at least find it.
    
    Regards, Marcel

Marcel,

  The number of times an action is fired is not really relevent
  to management, what they are interested in in the number of
  times specific events have happened. I would suggest that
  any customer wanting this information should write an action
  routine which receives ALL events and stores them in a database
  (such as RDB) so that they can then be reported on with one of
  the sophisticated report generators available in the market place.

  We could write the action to store the events, but, which database
  would we use, RDB? ORACLE?

Cheers,
Phil

    Marcel & Phil,
    
    OSCint V2.0 has a history action routine that stores all the events in
    a file (Not sure if it is RMS or RDB/Oracle).
    
    Where would we store it? - why in an Oracle database on a Quantum Disk
    drive of course.
    
    Martin