Did you resolve this issue? Something similar stumped us for a while as an error message was popping up a Windows alert box, and locking the condor job mysteriously until we had the "ah-ha" moment.. You absolutely have to get rid of that pop-up.
On Jun 4, 2009, at 7:50 AM, Michael O'Donnell wrote:
I am running Condor in a Windos 32 bit environment with about 6 machines added to the pool and one host (testing phase).
I have a program that apparently was developed in fortran and then compiled for a windows 32bit system. The program will run by passing a file input, but at the completion of the program the following popup occurs: "Program terminated with exit code 0". This code is generally an indication that a program ran successfully. Based on what I found online, this is an artifact of the compiler.
The problem: condor_submit will run the program but then then job never finishes, which I believe it is related to this popup. The GUI is suppressed on the individual machines by condor, all data is created and the CPU drops down to 0. The host reports that the job is still running and therefore, the data is not returned and the machine is tied up.