Reports Terminating Abruptly

[Nancy Palmquist]

George Simon wrote:
> I've seen this happen before and recently someone, maybe Richard K.,
> reported seeing the same behavior.
> I have a customer that, when running reports in filePro, will have the
> program exit abruptly, before reaching the 0 in the record count.
> The reports will exit at any point, not after processing the same number of
> records.  For example, a report that has selected 1,000 records will exit
> after processing only 20 or 600 or 950 but it will never run to conclusion.
> After the program exists the report will print normally but will not contain
> the data in the records not processed.
> This seems to happen mostly when a report selects a substantial amount of
> records.  The users I've seen this happen to, are always running (so far)
> Windows 2003 Server.
> I can download the user's data and run it in my system (XP or Vista) and the
> reports will run to their normal conclusion.
> Again, this does not happen to all my customers and not to all those running
> Windows 2003 Server.  So far I've had 3 customers that this has happened to.
> I lost one of those customer because of it.
> Does anyone have any idea what can be causing this to happen?
>
> George Simon
>
> _______________________________________________
> Filepro-list mailing list
> Filepro-list at lists.celestial.com
> http://mailman.celestial.com/mailman/listinfo/filepro-list
>
>   
George,

I have a similar chronic issue with a similar outcome.

A posting process that opens many many files will crash on Windows 
systems after roughly 100 records of processing.

The records are marked posted that it gets past, the rest are still waiting.

If the process is run again, it will pick up where it left off and 
continue.  It always seems to be between 80 and 120 records.  Most often 
it is less so most customers don't notice, but I have no access to the 
processing table on this report, so I can not see what is causing the issue.

Now I know this uses the most number of opens of anything we run and I 
am not sure about Writes or Closes being properly included.

I have also seen issues like this on a step that writes an index entry 
that it finds to be corrupt, it will crash without reporting an error.

I never have these kinds of issues on Unix/Linux systems.

Good luck.

check the following:

Indexes are not built on a bunch of blank fields - try to make entries 
as unique as possible.

Slow down the processing with a few sleeps after the file writes.

Make sure the output is sent to a file before Windows picks it up to 
print (a setting on the windows printer)

Make sure you have properly locked, and written stuff.

(Run it with PFCHECKLOCK=ON to see if your programming needs help)

Nancy

-- 
Nancy Palmquist 		MOS & filePro Training Available
Virtual Software Systems	Web Based Training and Consulting	
PHONE: (412) 835-9417		   Web site:  http://www.vss3.com