I think that the documentation and the product function are out of sync.  In
my opinion, MLM data collection/thresholding should work exactly like
NetView does, and convert COUNTER MIBs to a rate/second.  Two years ago, we
tried to deploy thresholding with MLM on version 5.0.0, and ran into this
problem.  At that time, MLM was not doing any delta calculations.  We got an
APAR opened, and the problem was partially fixed in 5.0.2.  I say partially
because I do not think the delta of the time values was ever added.  As a
possible workaround, you could change your threshold policy to divide the
MIBs by the number of seconds in your polling interval, e.g.
.1.3.6.1.2.1.2.1.1.10.* / 300 for a 5 minute polling interval.  This was
actually suggested to us by a L2 Support Engineer.  It's not a great
solution, but it does work.

In then end, we decided to use NetView for thresholding instead of MLM.

Joel Gerber - I/T Networking Professional - USAA

"I can please only one person per day.  Today is not your day.  Tomorrow's
not looking good either."
Dilbert's Words of Wisdom

        -----Original Message-----
        From:   Jane Curry [SMTP:jane.curry AT SKILLS-1ST.CO DOT UK]
        Sent:   Wednesday, October 13, 1999 15:18
        To:     NV-L AT UCSBVM.UCSB DOT EDU
        Subject:        MLM thresholding

        I have NetView 5.1.1 (still waiting for 5.1.2) and MLM 5.0.3.0 (from
NV
        5.1.1 CD).  I am using APM to setup thresholding.  So far, I am
simply
        sampling octets in (.1.3.6.1.2.1.2.1.1.10.*) on my 3 interfaces of a
        router.  Page 9-65 of the MLM manual talks about the Threshold
results
        table and says "Counter MIB variables default to the delta value,
        calculated by subtracting the old MIB value from the current MIB
        value".  Page 9-69, talking about Data Collection, says "threshold
for
        Counter type variables are computed by calculating the change per
second
        in the sampled values and checking these delta values against the
        threshold or rearm value".

        The thresholds I am seeing are firing on straight deltas (ie. the
P9-65)
        definition, NOT on the rate of change per second suggested on P9-69.
I
        have also tried another simple variable that just samples one
instance
        rather than several to ensure this isn't confusing matters.  I have
        checked with the MLM Threshold Results Table and that is confirming
what
        I am seeing.

        Do I have a code bug, a documentation bug, or am I just confused???

        Any help gratefully received,
        Jane
        --
        Tivoli Certified Enterprise Consultant
        Skills 1st Limited, 2 Cedar Chase, Taplow, Bucks, SL6 0EU, UK
        Tel: +44 (0)1628 782565
        Copyright (c) 1999 Jane Curry <jane.curry AT skills-1st.co DOT uk>.  All
rights
        reserved.