Conference azur::mcc

    
	Richard,

	You did everything correctly except one thing: you did not record 
  enough information to calculate past time statistics. As you know record
  operation allows to record every partition but statistics. So when Exporter
  is asked to perform a past time export it tries to calculate statistics
  on the fly from historical data. Node4 partitions do not provide all
  information needed for Performance Analyzer (PA) in order to calculate 
  statistics. As far as I remember you also need to record line characteristics
  and circuit counters (?). Probably the best thing is to look at Performance
  Analyzer Use manual. In this manual you can find a chapter which lists
  everything what you need to record. 

	Now about "time_stamp" field in the exported table. This field is 
  not used by the report package. Under different circumstances this field can
  have different meaning. But  in each table you can find "time_<partition_name>"
  field for each partition (for example, "time_counters"). This information
  together with status ("mcc_cvr_<partition>) gives you real time stamp and
  status of the original reqiest. As far as I remember Historical Data Use
  manual has a description of the data base table structure.

	Sam
  
 

    

    

    
    Whoops, 
    
    	sorry, but I guess I missed it in my note. Yes, I did record line
    statistics.. I recorded line, node, and circuit statistics for the
    node, and still I got no data.
    
    	Anyways, I will try yet again to get mcc working properly, and will
    report back.  Maybe I missed it.
    

No, please check Sam's note again.  You can not record statistics at all.
You must record the raw data required by the statistical calculations.
Then when you try to review (or export) statistical data, it is
re-computed from the historical raw data.

-- Erik

    
    Ok, I have partial success here. I am able to run reports on exported
    data, and the data is valid.
    
    I still don't see anything different this time, then what I did last
    time. (when I said recording statistics in .2, I shouldn't have used
    the word statistics, I was using statistics as a general term for
    characteristics, counters, etc..)
    
    Anyways, the data is being recorded, and the "time_stamp" column is now
    incrementing properly. (again, don't ask me why it is now, when it
    didn't in .0. I'm using the same commands).
    
    I am now getting plaged by errors of the following time when I try to
    export data (again, only sometimes, and only on some items.) Note that
    TROR01 CIRCUIT SYN-0 is a registered object, and I am executing this
    command at approx 16:00 hrs.
    
    MCC> EXPORT NODE4 TROR01 CIRCUIT SYN-0 -
    _MCC> EXPORT TARGET=DISK$DATA1:[REPORTS]REPORTS.RDB, -
    _MCC> EXPORT PERIOD=00:15:00, -
    _MCC> BEGIN TIME=TODAY, -
    _MCC> END TIME=NOW, -
    _MCC> IN DOMAIN .TRO.S.MCC.RECORD
    
    
    The requested operation cannot be completed
           MCC Routine Error = %MCC-S-NORMAL, normal successful completion,
           Entity Existence Info = Entity Existence Cannot Be Determined
    
    
    But here is the silly thing. It did export some data, as is shown by
    the SQL command below
    
    NODE4      TIME_STAMP
     57.56      29-AUG-1993 23:40:33.74
     57.56      30-AUG-1993 00:10:41.07
     57.56      30-AUG-1993 00:10:41.07
     57.56      30-AUG-1993 00:40:33.96
     57.56      30-AUG-1993 00:50:38.05
     57.56      30-AUG-1993 01:10:40.87
     57.56      30-AUG-1993 01:20:36.74
     57.56      30-AUG-1993 01:40:48.41
     57.56      30-AUG-1993 01:50:39.88
     57.56      30-AUG-1993 02:10:41.04
     57.56      30-AUG-1993 02:20:45.05
     57.56      30-AUG-1993 02:40:39.85
     57.56      30-AUG-1993 02:40:39.85
     57.56      30-AUG-1993 03:10:42.14
     57.56      30-AUG-1993 03:20:38.07
     57.56      30-AUG-1993 03:40:37.37
     57.56      30-AUG-1993 03:50:41.52
     57.56      30-AUG-1993 04:10:40.90
     57.56      30-AUG-1993 04:20:46.48
     57.56      30-AUG-1993 04:40:44.98
     57.56      30-AUG-1993 04:50:39.48
     57.56      30-AUG-1993 05:10:49.15
     57.56      30-AUG-1993 05:20:38.75
     57.56      30-AUG-1993 05:20:38.75
     57.56      30-AUG-1993 05:20:38.75
     57.56      30-AUG-1993 05:20:38.75
     57.56      30-AUG-1993 06:20:37.36
     57.56      30-AUG-1993 06:20:37.36
    28 rows selected
    
    
    
    So, the question I ask is, why did it export some of the data, and then
    crap out with this error, when It was able to determine the existance
    before it died?  Please don't tell me that the code keeps checking, and
    checking, and checking in dns every single time!  Is this why exporting 
    from the historian is so slow?
                                
    BTW, again I'm using mcc1.3 and VMS 5.5-2, using DNS.

    
    ok folks, now I'm really getting upset with MCC (-:
    
    here are two instances of the same line of code, with just a slight
    change, causing a syntax error.  I point out, that it's the exact same
    piece of code in the DCL procedure both times.  Here is a cut and paste
    of the two commands.
    
    	Note that these are cut and pasted directly from the log of the 
    command procedure.  The only difference is the Nodename
    
    $               manage/enterprise -
                            export node4 KAOR03  -
                            export target disk$data1:[reports]reports.rdb, -
                             export period=00:15:00, -
                                    begin time = 30-aug-1993 00:00:00.00, -
                                    end time = now, -
                                    in domain .tro.s.mcc.record
    DECmcc (V1.3.0)
    
    
    Node4 DEC:.kao.kaor03
    AT 30-AUG-1993 17:28:27
    
    Exporting completed
    
    
---
    $               manage/enterprise -
                            export node4 KAOR04  -
                            export target disk$data1:[reports]reports.rdb, -
                            export period=00:15:00, -
                                    begin time = 30-aug-1993 00:00:00.00, -
                                    end time = now, -
                                    in domain .tro.s.mcc.record
    DECmcc (V1.3.0)
    
    %MCC-I-NOPARCMD, .TRO.S.MCC.RECORD
                     ^
    %MCC-I-SYNTAXERR, Syntax error -- unable to interpret remainder of line
    %MCC-S-NORMAL, normal successful completion
    
    ------
    I can't find any reason to this. There is no discernable difference between
    the two commands.   This command is in a simple subroutine where p1 is
    the node to export.
    
    Here is the simple command procedure that I wrote that will automatically
    export the data from my historian to the rdb file. It first uses SQL
    to look into the database, to find out where the last export crapped out 
    (or ended), and continues where it left off. I wrote this to help work 
    around the problem described in the previous reply. 
    
$ SET NOON
$ SET VERIFY
$! router TROR01
$!	This procedure should be called periodically to export 
$!	node4 data to the rdb file disk$data1:[reports]reports.rdb
$!	The procedure looks into the rdb database using SQL to determine
$!	when the last export completed.  It then exports from that time
$! 	to the present time using an interval of 15 minutes.
$
$ set default disk$data1:[reports]
$
$ call export TROR01
$ call export TROR02
$ call export KAOR02
$ call export KAOR03
$ call export KAOR04
$ call export CTHR01
$ call export OTOR01
$ call export MQOR01
$ call export WNOR01
$ call export EMOR01
$ call export VAOR01
$ call export CGOR01
$
$ exit
$
$ export:
$	subroutine
$
$!	p1 - name of the entity to export
$
$! first to get the maximum date from the sql database
$! for the node p1
$
$	filename = "select.tmp"
$	quote :="'"
$
$! write the sql to find the maximum date
$!
$	open/write file$ 'filename'
$ 	say := write file$
$	say "$ mcr sql$                                             
$	say "	attach ",quote,"alias d filename disk$data1:[reports]reports.rdb",quote,";"
$	say "	select max(time_stamp) from d.mcc_node4 where
$	say "		name containing ",quote,p1,quote,";"
$	say "$ exit"
$	close file$
$
$! execute the file created
$! note that verify must be off so that the contents of the file
$! is known. this file must be parsed later.
$!
$	v=f$verify(0)
$	@select.tmp/out=select.tmp_out
$	v=f$verify(1)
$!
$! see if a row was selected
$!
$	search select.tmp_out "1 row selected"/output=nl:
$!
$! one row was selected, so hopefully this was a date
$!
$	if $severity=1 
$	then
$!
$!		get date from the file
$!
$		open/read file$ select.tmp_out
$		read file$ line
$		read file$ line
$		startdate == f$edit(line,"trim")
$		close file$
$		write sys$output startdate
$!
$!		there were no records for this node in the rdb database
$!		so assign a startdate manually
$!
$		if startdate .eqs. "NULL"
$		then
$			startdate == "30-aug-1993 00:00:00.00"
$		endif
$
$!		execute the export
$		manage/enterprise -
			export node4 'p1'  -
			export target disk$data1:[reports]reports.rdb, -
			export period=00:15:00, -
				begin time = 'startdate', -
				end time = now, -
				in domain .tro.s.mcc.record
$	else
$!
$!		nothing was selected. Something is messed up so save 
$!		the output, and quit
$!
$		copy select.tmp_out select.tmp_bad
$	endif
$
$	delete select.tmp;*
$	delete select.tmp_out;*						
$
$ endsubroutine