Description of problem: This bug shows the same charateristics as Bug 136657 ( https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=136657 )but in Red Hat ES3 update 7 it has an additional side effect. The additonal side effect is described in this scenerio.. 1. When crond runs /usr/lib/sa/sa1 from /etc/cron.d/sysstat, the following error is logged into /var/log/cron Apr 4 20:00:01 rh crond[4512]: (root) CMD (/usr/lib/sa/sa1 1 1) Apr 4 20:00:01 rh crond[4512]: LAuS error - do_command.c:278 - laus_attach W : (19) laus_attach: No such device 2. A user edits his crontab and saves the changes 3. The crond deamon for some reason is unable to reload changed user crontab entries and still runs the old entries (No reload message in /var/log/cron) I run crond with debug on ( # crond -x load ) debug message indicates that the spool directory mtime is not updated, which should not be the case Version-Release number of selected component (if applicable): vixie-cron-4.1-10.EL3 sysstat-5.0.5-11.rhel3 How reproducible: Easy Steps to Reproduce: 1. Install sysstat-5.0.5-11.rhel3 2. Stop the linux Audit subsystem (/etc/init.d/audit stop) 3. Ask user to edit and save crontab settings Actual results: 1. User's new crontab settings not implemented Expected results: 2. Users's new crontab settings are implemented Additional info: 1. This problem does not show up in Red Hat Advanced Server 3 update 6 even with audit off and sysstat installed
You need to reboot the machine after disabling audit: # service audit stop # chkconfig --del audit # reboot Once rebooted, these problems do not occur - see bug 177321. I've seen no problems with crontab updates not taking effect - they may take up to 2 minutes for crond to reload them.