We have a nightly job that as part of it's duties is to:
Seq 30 Release to Invoicing (batch)
seq 40 Global Confirmation of future invoices
Seq 50/60/70 Compose Print & Post for 3 financial companies

for reasons best known to itself, about 2 weeks ago Seq 50 did not find any invoices to print. After several trials and errors have the following work around

Seq 50 has been copied into Seq 55 and both are active - when job runs, seq 50 still does not find invoices but seq 55 does. If seq 50 inactive, seq 55 fails.

there are no error messages, the sessions are running 'correctly' but completely baffled (as are Infor support) as to exactly what has caused this and what is going on

Any contributions gladly welcome

The first thing that crosses my mind is that maybe seq 60 is starting too fast, i.e. too short after 40 is done and therefor the data is not yet available.

Therefor when you run it again after failing, and thus some time has passed, the data will be found, and the process will success.

For the same reason, obviously if you remove seq50, then seq55 will fail for there again is no break between 40 and 50/55.

It is just a thought, but you can easily test it. I don't know if your programs are the standard Baan sessions or customized scripts, anyway if customized and you own sources, then you can just write suspend(30000) in the before.program, to wait 30 seconds before continuing. If you don't ahve access to sources, you can maybe do this through a batch file. Again depending on the task schedular you use, this is easily done..

Let us know your findings please..


Good luck,
Eli Nager

That was one of the many attempts at a fix - relocated another session from the job that runs for about 2 or 3 minutes before sequence 50, made no difference, but thanks anway for the suggestion

Are seq 50,55,60 & 70 the standard Baan sessions? If not, can you please send us your code...

All are standard Baan - we do not have code and aim for as minimum customisations as possible

Have you deleted the sequence 50 and re-added from the session? Does it work if you start the job manually? You probably already tried these, but I thought I would ask. I have seen the batch parameters get corrupted for a job. We had to re-create the sequence from scratch - only happened once a couple of years ago.

Have you applied any patches lately? Are reports involved with 40 and 50. If so what kind of devices are you using?

Wow a second thought - wooohooo 2 in one day. If when you activate it manually it does not work get Baan to give you a debug object. Then you can launch in manually and see what is not working the first time and what is working the second time throught. I just ran into a situation where the first time a program ran it did not know about a global variable used in a library and yet the second time through it new about it. I think this may be something along these lines. Of course I know nada about Baan 5.

Mark,

I have myself come across this phenomenon of non-recognized variables/values during sessions' first execution. But the cases I saw, where all when a session is executed twice WITHOUT closing the session form.

David's case is a session which is executed, closed, and started again... Could this phenomenon happen in this case as well..?? I'm interested to know...

Regards,
Eli

I was thinking about my post in the AFS forum on spool.fileout. My program opens tisfc9475m000 opens tisfc9417m000 the first time with temp file. The program tisfc9417m000 generates a report, and then closes. So you would think everything would clear out. Nope the second time the session tisfc9417m000 runs it knows about the first temp file name and not the second. When I have them both in debug spool.fileout is correct tisfc9475m000, but for some reason tisfc9417m000 gets the old value. How I closed the session and have waited upto 30 minutes with tisfc9475m000 open to run it again. I am just wondering if it could be the same type situation. I think a debug mode object would help solve the problem and I am really curious as to what might cause the problem.

Thanks for the input - further update. Created an entirely new job using these sessions in the same sequence (did not create seq 55). Same result, the first compose, print post session found no data, the nexr two did

One question I did not see answered - if you log into the system and activate the job from your logon - does it still fail?

Run manually, job works fine