Description of problem: After submission ~1.2M jobs has scheduler in memory almost 3GB. I think after discussion with developers that It could be possible optimize this huge memory allocation. Very simple example of job: Universe=vanilla Executable=/bin/sleep Arguments=10 Queue 1000000 Version-Release number of selected component (if applicable): condor-7.4.3-0.19.el5 Please feel free to change target milestone.
Here are examples that scheduler crash because it reach 3GB of memory which is border of OS with PAE kernel. Please see #603657 #603658
*** This bug has been marked as a duplicate of bug 742874 ***