Conference acadmy::tpas_notes

	Lois,

	The following document will be forwarded to ACADMY::QCA_ADMIN ...

	TPAS DATABASE EXTRACT FILES

		Leslie and I both agreed that a mail message should be sent 
		only when there is an exception to the availability times.  
		Past experience dictates that scheduled messages end up being 
		deleted without being read.

		When an exception to the availability of the TPAS RDB extract
		is encountered, the production processes can be re-scheduled.
		This helps encourage exception processing and lights out
		environment which many of the manufacturing sites are working 
		towards.
	
		If the change report was mailed with the specified information
		(file sequence number), then the mail message would be more
		practical and it would be read.

		We do not have any timing issues around this extract.  My only
		concern is if a plant discover a problem with the data.  The 
		Thursday 4:00 availability does not give the plants a lot of 
		reaction time.  I think this is a Corporate Tax issue that 
		should be addressed.


	TPAS STOCKROOM DATABASE EXTRACT FILES

		We definitely have a timing issue with the availability of
		the Stockroom file.  We are aware of the fact that this is
		not a rapidly changing file however with the introduction of
		TPAS into manufacturing our day-to-day operations have now
		become dependent on the stockroom file.

		Our monthend process begins at midnight and we are striving to 
		have our systems available to manufacturing for Sunday morning 
		at 8:00.  If we encounter any problems while retrieving this 
		file we do not have a lot of time to recover our processes.  
		We are setting ourselves up for failure.  Networks can sometimes
		be unreliable (during storms, shut downs and crashes).  How 
		would we recover our processes if we are not able to retrieve 
		the stockroom file.  If possible, I would like to ask your group
		to re-evaluate your processes in order to provide the Stockroom 
		files to the manufacturing plants in a more timely fashion.

		This would be greatly appreciated.  


	ONE LAST POINT 

		Since the TPAS import processes are running within a dec 
		scheduler environment, the mail messages sent can not restrict
		or execute the TPAS import processes.  Do you have any 
		recommendation for how to ensure we have the most current files.

		If a file is reworked and made available to the plants, how
		can we determine if processes are working with the most 
		up-to-date files.  Don't forget, the TPAS file are created
		during monthend ... some forecast reports require the new files.
		(Kanata import pricebook used by Northboro for importing into
		Canada is one example).


	Sincerely,

	Bridget Terpstra	

>	Since the TPAS import processes are running within a dec 
>	scheduler environment, the mail messages sent can not restrict
>	or execute the TPAS import processes.  Do you have any 
>	recommendation for how to ensure we have the most current files.
>
>	If a file is reworked and made available to the plants, how
>	can we determine if processes are working with the most 
>	up-to-date files.  

    We have the same problem. It is basically impossible now to determine
    up front, which RBR file we should be waiting for. What we do now is
    that on Sunday evening at 8 o'clock we copy all RBR file from the
    TPAS$OUTBOUND area, regardless of whatever creation date or sequence
    number they have. Next we will process the RBR files with the highest
    sequence number. According to the current schedule the proper files
    should be available by that time, but this method is of course a bit
    unreliable (e.g. if for some reason the new RBR files are not available
    at 8 o'clock on Sunday evening, we'll just copy the old ones (which
    means we will NEVER process the proper new RBR files)).

    I think the only way to solve all this is by somehow making the
    monthend RBR files uniquely identifiable. This way we are always able
    to determine what file we should be waiting for. Possible ways to do
    this are:
    - use a sequence number per month (this means if a special RBR file
      is created somewhere in between two monthend's, give it the same
      sequence number as the (regular) one from last monthend). This way we
      can look for a file with a specific sequential sequence number.
    - a couple of days before monthend remove all RBR files from
      TPAS$OUTBOUND. This way we can wait until a RBR file appears and copy
      that one.
    - add something to the name of the RBR file identifying the month, e.g.
      TPAS_STKRM_DB_0033_APR_1993.RBR. - etc.

    Personally I prefer the first option.

    Besides better reliability and controllability an other advantage of
    uniquely identifiable files is that we could write a process which will
    wait for the required file, pick it up and process it immediately 

    Is there any chance of getting this mechanism (or something similar)
    implemented? Does anyone have a smarter way of copying the proper RBR
    files at monthend?

Regards,
    Andr� van 't Hoog