[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference azur::mcc

Title:DECmcc user notes file. Does not replace IPMT.
Notice:Use IPMT for problems. Newsletter location in note 6187
Moderator:TAEC::BEROUD
Created:Mon Aug 21 1989
Last Modified:Wed Jun 04 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:6497
Total number of notes:27359

1709.0. "Changing default duration " by TOOK::ANWARUDDIN (Anwar) Thu Oct 24 1991 15:37






      With this note I am soliciting your opinion  about  the  following

      change we would like to make in DECmcc V1.2 PA.



       Currently, DECmcc PA assumes a DEFAULT DURATION of 5 minutes over

      which   to  calculate  the  statistics  if  the  duration  is  not

      specified.  For online (realtime) statistics it meant waiting  for

      a  period  of  5 mins.  before you could see the statistics on the

      screen.  And when you are used to seeing  other  attributes  being

      displayed  a  very  short duration after the request is made, this

      behavior leads one to assume that the system is hung.  If you  are

      using realtime graphing capability - available in V1.2, to monitor

      the performance, the 5 minute delay could be  annoying.   So  with

      DECmcc  V1.2  PA we would like to change the DEFAULT DURATION to 1

      minute.



       What do you think?

T.RTitleUserPersonal
Name
DateLines
1709.1DO IT !!!CHRISB::BRIENENDECmcc Bridge|Station|SNMP Management.Thu Oct 24 1991 15:380
1709.2Please DO IT, but document it clearly in the RELEASE_NOTES, and Userguide for PAVIDOAN::DOANEIS/Network Consulting Group of the 90'sThu Oct 24 1991 15:5011
Hi,

	I support the default duration of 1 minute, but another thing that
	is REALLY needed is the CONSISTENCY in specifying the time parameters
	while using the PA because it's very difficult ( impossible !!) to
	memorize the complex syntax that we're having in V1.1 (especially for
	retrieving the historical data from the MIR to get statistics.


Thien Doan
Network Consulting Group
1709.3inform the userJETSAM::WOODCOCKThu Oct 24 1991 17:0918
Hi,

I really don't have a strong preference for either but 5 minutes seems a
bit much for a generic default for a user to have to WAIT on. Whatever
you decide though, how about TELLING the user what is being done:

Ex.

MCC> show node4 XXXX all statis

	Using default duration of 1 minute...



just a thought,
brad...

 
1709.4We can do this even betterTNPUBS::JONGSteve Jong/T and N PublicationsThu Oct 24 1991 21:294
    I've said it before and I'll say it again: an in-progress message is
    necessary here.  Simplest would be a box analogous to the "Looking
    into" box that would say "Gathering statistics for ..."  Better would
    be a countdown timer ("Statistics due in n:nn minutes")
1709.5GRANMA::JPARKERThu Oct 24 1991 22:475
    RE: 1 minute duration. Yes, do it. I usually set a lower duration
    anyway. In either case, a status message is very much desired.
    
    Jeff.
    
1709.6Go for 1 minute default + provide "count-down"CUJO::HILLDan Hill-Net.Mgt.-Customer ResidentFri Oct 25 1991 01:228
    I like Steve's idea of the count-down timer, if you have time to code
    something like this for V1.2.  I'm also assuming that we will still be
    able to "CANCEL" if desired.
    
    In-progress messages, pointer "watches", and other such amenities make
    using DECmcc much more of a pleasure.
    
    -Dan
1709.7and not just for PA, please?ASD::HOWERHelen HowerFri Oct 25 1991 11:255
If countdown timers or other 'work in progress, please wait' type flags are
implemented, they should be made generic if at all possible.  Other FMs (and 
maybe AMs?) could use the ability to display such flags during long operations 
so the user knows they're still alive! :-)
		Helen
1709.8TOOK::STRUTTManagement - the one word oxymoronFri Oct 25 1991 19:015
    How about a default of 10seconds.
    I can wait 10 seconds - longer and I get bored or assume the machine's
    failed....
    
    Colin
1709.9ENUF::GASSMANMon Oct 28 1991 12:3714
    It depends on who the consumer of the stats are.  When doing some
    real-time tuning, a refresh rate of 1 second is not unreasonable.
    A 10 second refresh is good for monitoring things with rapid rates
    of change, such as bandwidth utilization and number of active session
    links - but for everyday background stats to glance at once in a while,
    the 1 minute default is probably good.  The countdown visual idea is
    good.  Another helpful feature might be some 'standard value' buttons
    so the user can click on 1, 10, or 60 seconds right from the display.
    The MCC concept of shuting down if the interval can't be met (ie,
    polling counters on a long delay link) should be modified to allow
    stats to be updated on a 'best available' basis.  
    
    bill
    
1709.10Simplify itTOOK::R_SPENCENets don't fail me now...Tue Nov 12 1991 17:1811
    I like Bill's suggestion of providing a very simply way of getting
    a few common durations so that the existing operations time ....
    can be avoided most of the time.
    
    I also believe that (while we are talking about time and statistics)
    when a user does a OPERATIONS SHOW STATISTICS, the system ought
    to know that it has to use relative time and not complain to the
    user to fix it. Then they have to fix it back for the next operation
    (usually SHOW COUNTERS).
    
    s/rob