That memory usage does not look unreasonable.
Please try to capture similar output after you've seen the schedd grow
to 900MB with no jobs.
Best,
OK will do but as ever it looks difficult to reproduce. What I can see
already is that I've doubled to memory use to around ~ 70 MB since
yesterday through submitting new jobs and the pool state is currently
pretty much the same as beforehand when the memory use was ~ 35 MB.
I still don't see how this fits in with the documentation which claims that
the schedd uses about 10 kB for each job. Is the documentation out
of date ? In which case what would a more accurate figure be.
many thanks,
-ian.