[Condor-users] standard universe jobs with lots of stderr output

[Stuart Anderson <anderson@xxxxxxxxxxxxxxxx>]

For standard universe jobs with error defaulting (or explicitly being set) to /dev/null which /dev is used, the one on the submit or the execute?

Note, this is a scalability question with respect to a large number of shadow processes running on a single submit machine.

Thanks.

--
Stuart Anderson  anderson@xxxxxxxxxxxxxxxx
http://www.ligo.caltech.edu/~anderson