+++ This bug was initially created as a clone of Bug #129920 +++ Description of problem: In SE Linux the program that is executed determines the domain that the program is run in. Running "/usr/bin/python -S /usr/lib/mailman/cron/cull_bad_shunt" from /etc/cron.d/mailman causes the cron job to run in domain sysadm_crond_t, while running it as "/usr/lib/mailman/cron/cull_bad_shunt" (with the first line of the script being "#!/usr/bin/python -S") causes it to run in domain mailman_queue_t (really this one? Jan K.) which is the desired result (*). This change will not have any affect for people who don't use SE Linux and will do a lot of good for SE Linux users. Version-Release number of selected component (if applicable): mailman-2.1.11-2.fc10.x86_64 (this line was not present in mailman-2.1.9-10.fc9.x86_64) How reproducible: Always Steps to Reproduce: examine /etc/cron.d/mailman Additional info: I do not run SELinux so I know nothing if it is a problem. Just IMO it is as I saw the old bug and the current file content. Sorry if it works in SELinux.
You're right. This doesn't seem to be OK. I didn't see any AVC messages caused by this yet but the shunt and bad queues are usually empty on my system... Thanks.
Mailman-2.1.11-3 should solve the problem. Please reopen the bug if not.