Description of problem: The man page of rhns says: rhnsd is typically started from the initialization (init) scripts in /etc/init.d/rhnsd. when it’s time to poll the Red Hat Network servers for available updates and actions. The default interval is every 240 minutes. The minimum polling interval is 60 minutes. and Specify the interval that rhnsd should wait between checking the Red Hat Network. Default is 240 minutes, the minimum is 60 minutes. However, when I change the default INTERVAL=240 in /etc/sysconfig/rhn/rhnsd to INTERVAL=40 I get the following message in /var/log/messages: rhnsd: Red Hat Network Services Daemon running with check_in interval set to 40 seconds. Therefore, the problems are: 1) 40 seconds < 60 minutes which is the claimed minimum. 2) Since the default value is 240, it means that the value is in minutes, not in seconds as the message in /var/log/messages has us believe. Version-Release number of selected component (if applicable): rhnsd-4.8.3-1.el5 How reproducible: Deterministic. Steps to Reproduce: 1. With default configuration (240), run service rhnsd restart and watch /var/log/messages. 2. Change the value in /etc/sysconfig/rhn/rhnsd to 40 and run service rhnsd restart; watch /var/log/messages. Actual results: Feb 9 07:13:35 hp-dx2200-01 rhnsd: Red Hat Network Services Daemon running with check_in interval set to 240 seconds. Feb 9 07:13:35 hp-dx2200-01 rhnsd[7029]: Red Hat Network Services Daemon starting up. Feb 9 07:36:57 hp-dx2200-01 rhnsd: Red Hat Network Services Daemon running with check_in interval set to 40 seconds. Feb 9 07:36:57 hp-dx2200-01 rhnsd[7266]: Red Hat Network Services Daemon starting up. Expected results: Red Hat Network Services Daemon running with check_in interval set to 240 minutes. ^^ Note minutes, not seconds. Red Hat Network Services Daemon running with check_in interval set to 60 minutes. ^^ Note that the 40 minutes in config value was adjusted to the documented minimum, 60 minutes. Additional info: This bugzilla was prompted by bug 533838 (private).
Taking.
Actually, another issue is that the "running with check_in interval" message is always printed twice: Feb 9 07:13:35 hp-dx2200-01 rhnsd: Red Hat Network Services Daemon running with check_in interval set to 240 seconds. Feb 9 07:13:35 hp-dx2200-01 rhnsd: Red Hat Network Services Daemon running with check_in interval set to 240 seconds. Feb 9 07:13:35 hp-dx2200-01 rhnsd[7029]: Red Hat Network Services Daemon starting up. Feb 9 07:36:57 hp-dx2200-01 rhnsd[7029]: Exiting Feb 9 07:36:57 hp-dx2200-01 rhnsd: Red Hat Network Services Daemon running with check_in interval set to 40 seconds. Feb 9 07:36:57 hp-dx2200-01 rhnsd: Red Hat Network Services Daemon running with check_in interval set to 40 seconds. Feb 9 07:36:57 hp-dx2200-01 rhnsd[7266]: Red Hat Network Services Daemon starting up. That should also be fixed.
Another issue: if there is /etc/sysconfig/rhn/rhnsd with INTERVAL=240 and rhnsd is started with --interval 80, it will run with value 80. If the daemon is then sent HUP, it will run with 240, not 80 set with the command line option. Is that correct?
Fixes in Spacewalk master 5fe3abb16874bb8ac34706883a0530cbe1e8e2e2 to 51f27ff6e5e01df1023a07fe52660fc25dd1d237.
Note that we will likely want to remove the actual handling from SIGHUP_handler and just set some flag and alarm the sleep as well.
(In reply to comment #5) > Note that we will likely want to remove the actual handling from > SIGHUP_handler and just set some flag and alarm the sleep as well. Actually, it seems to behave reliably.
Tagged and built as rhnsd-4.9.3-1.
The fix was part of Spacewalk 0.8 client release.
*** Bug 554756 has been marked as a duplicate of this bug. ***