Re: We have a mystery

Darren, I've run into this even without SQL when I have a file defined for output.  Sometimes things get hung up in the cache.  The old school approach was to compile the file with or override it to FRCRATIO(1), even older school was make the file both input and output (and back then you had to have a dummy read or the compile complained).

Nowadays I use FEOD.  It's a nice little opcode.

But in nearly all these cases, it's because the program writing the file still has the file open while I'm trying to do the read. Is the program that writes ending before the SQL is invoked?

We have an issue with a vendor system writing data to a file using
traditional RPG, and then opening a cursor against that file pretty
quickly afterwords.  Intermitently, the cursor finds no data, but if I
have it retry it sometimes works.  This began after we applied some PTF's
in March 2017 to our i7.3 system.  Neither IBM nor the vendor have been
able to reproduce the issue, so it continues.  Through the vendor, we
found out another user with a similar system to our is also having this
issue.  I'm not saying its the cause of yours, but I'd be really
interested if you found a solution and to see if it applies to us.

___________________________________
Darren Strong
Dekko