From Bugzilla Helper: User-Agent: Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1) Description of problem: The restart option of the /etc/init.d/named script executes stop and then start (immediately after it). The named isn't starting if there is still a 'named-process', which is the case in the first few seconds after the stop command has been given in the case of a heavy loaded DNS server. Because of that there is a risk that the restart doesn't really do a restart and returns code 1. I think the solution is to alter the script like this: restart() { stop sleep 3 start } Version-Release number of selected component (if applicable): bind-9.2.1-9 How reproducible: Sometimes Steps to Reproduce: 1. execute /etc/named/restart on a highly loaded DNS server 2. 3. Actual Results: Output: named already running Return code 1, and named doesn't start Expected Results: restart of named Additional info: This problem can be severe in a case where you have 2 DNS servers, with a heartbeat configuration so that the 2nd DNS server takes over when the first goes down. As this 2nd takes over the IP address of the first, we configured heartbeat to issue a /etc/init.d/named restart, the named doesn't start and ... no DNS!
THe latest bind script currently has this fix. (Sleep 2)