Jerry,
1. 1st thing I'd do is look in the recent changes for a construct
functionally something like this:
X cleared (wrong place for X to be initialized.)
Start-outer-loop
* X should be cleared here, but isn't.
Start-inner-loop
X:= [stuff] (X grows)
end-inner-loop
End-inner-loop
You say it runs fine by itself, but not in it's rightful setting, so
maybe the var is in common and persists through multiple executions w/o
proper reinitialization.
2. I recently had a memory leak where executions got slower & slower,
until the uv process aborted. Different error message from yours.
Unix's "ps -l" was interesting. Size column showed it growing. It
turned out ULIMIT config param was set too high. Joel Yates @ IBM had me
reset it to the default. Hadn't been an issue until we changed from
HP-Risc/HPUX11.11/UV10.1.16 to HP-itanium/HPUX11.23/UV10.2.6.
Thank-you, Joel.
cds
-----Original Message-----
Last night during our month end batch process it aborted the proc when
we got "Available printer memory exceeded" during one of the programs. I
ran the program this morning and it ran without getting the error, same
program, same data files. Can anyone tell me where I could look to keep
it from doing this? It has always run without problems in the past, it
only runs at month end, but there were some minor changes made to the
program that shouldn't have caused any problems and don't when run by
itself.
Jerry Banker
-------
u2-users mailing list
[email protected]
To unsubscribe please visit http://listserver.u2ug.org/
