Just a thought, which may be a simple solution, but suiting most people.
The kernel is pretty good at killing misbehaving jobs. @Ole: can you
capture SIGKILL from a job? Can you record memory usage shortly after
SIGKILL?
If yes, you have an estimate on the minimum memory requirement of the job
an
Hi Sebastian
we do need swap to permit finishing of jobs in some
exceptional instances, and then lots of it. On the other hand, use of swap
should
be really avoided as it slows down jobs to a crawl, and eats up disks.
I am not convinced to go this route. i.e. let the kernel kill already
misbeha
