Conference azur::mcc

Hi,

Please forgive what may be a silly question.

Does this condition cause the rule to "fire" or to encounter an exception?

If the customer is currently using the same procedure for both his "procedure" 
and the "exception handler", perhaps it would be acceptable to simply
create a different procedure for the exception handler which would
be smart enough to ignore these spurious problems.  From the V11 online help:



ENTITY

  ALARMS

    RULE

      CREATE

        Exception_Handler

          Information


            When an error occurs during the evaluation of an alarm expression,
            the error is passed, as string values, to the command procedure
            through the following parameters:

              P1 - Rule name
              P2 - Description
              P3 - Category
              P4 - Expression
              P5 - Time of error detection (Binary Absolute Time output
                   format)
              P6 - Error that occurred.  This parameter will also contain
                   the string "The rule has been disabled." if the error caused
                   the rule to be automatically disabled. There may be a number
                   of separate text strings used to describe the error and/or
                   the state of the rule.  When this is the case, they will be
                   separated by the string "<EOS>" (End Of Segment).
              P7 - Parameter
              P8 - Data file.


ENTITY ALARMS RULE CREATE Exception_Handler Subtopic?

    Pete,
    
    The alarm rule fires an Exception, but the silly thing is that there
    should be no exception (neither a normal alarm) being fired if (an I am
    assuming this) the network is stable and operational.  If the cause of
    the exception is something related to the management system (And I
    presume this), that situation is not acceptable.
    
    If the problem is in the management station then,
    	if it is a resource problem, that should be solved.
    	if it is a DECmcc code problem, that should also be solved.
    	if it is e.g. because the management station does not see the
    network, that would be acceptable, but then there should be hard
    evidence (with DECnet EVL) that the station was temporarely
    "disconnected".  For this last assumption, no event has been seen
    around the intervals when the exception happens, although event
    logging has been sinked to the management station, and at other moments
    "normal" events are being collected.
    
    Therefore I want to investigate the cause of the "spurious" exceptions.
    
    Regards,
    	Dominique.
    

You might send/post an exact log of what's going on so the Phase IV folks can
be precise in figuring out what's going on?  Jean Lee is responsible for DNA4.

I looked into this.  No place in the AM do we explicitly set this
exception.  We truly returned from a QIO with SS$_UNREACHABLE and
translated it to the specialized exception which you are seeing.
As far as VMS was concerned, the node was unreachable.

We will continue to look into this problem but I figured I'd give
what info we have in the meantime.

	-- John

One possible explanation of this behavior is that the router is unable respond
to the management system's connection initiate message before the outgoing timer
on the management system expires. 

This type of timeout can be caused by a number of factors including: network
congestion and network routing cost transitions. Could you please check the
value of the outgoing timer and the number of response timeouts on the
management system and the value of the incoming timer on the router. Also can
you give me some idea of the topology between the management system and the
router.

Thanks,

Bill
    

    Thanks for specifying that the error "Node not currently accessible"
    unambiguisly corresponds with SS$_UNREACAHBLE (could have been
    documented !)
    
    Now it's also clear that DECmcc is out of suspicion.  I can reproduce the
    error (quite difficultly) with NCP procedures.  It looks at this time
    as a DECrouter/X25 gateway problem.  Investigation goes further.
    
    	Dominique.