Description of Problem:
If an error exists in the /etc/named.conf file, running the "service named
start" command, causes an "OK" message for starting the service. In
actuality, the service is not started, so it is not "OK".
Steps to Reproduce:
1. Stop named (type "service named stop" as root)
2. Open the previous working /etc/named.conf in an editor and add some
random text. Save the file.
3. Start named (type "service named start" as root)
You will see OK but the service is not running. The error is correctly
posted in /var/log/messages
The service is not started when the user thinks it is started.
The service should report back that it "Failed" letting the user know that
it is not running.
The problem is that the named parent process returns 0 even if the config
files are broken (because the config files are parsed later by the various
I've added a workaround to the startup scripts in 9.1.3-3.
I can confir this bug
Perhaps users should use named-checkconf to test their configurations instead of
re-starting the server? I sure wish I'd known about it when I upgraded from bind
8 to 9.
Yes both named-checkconf and named-checkzone are useful - but that is not the
bug here - this bug most often shows up in linuxconf thinking that named is not
running when it really is and tries to start a second copy.