From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.2.1) Gecko/20030225 Description of problem: I've got a shell script that rewrites /etc/named.conf to include new secondary zones. It then does 'rndc reload' to make them take effect. It used to work with older versions of bind, but not now. Version-Release number of selected component (if applicable): bind-9.2.1-9 How reproducible: Always Steps to Reproduce: 1. echo 'zone "foobar.com" {type slave; file "sec/foobar.com"; masters { 2.2.2.2; }; };' >> /etc/named.conf 2. rndc reload 3. grep foobar.com /var/log/messages 4. nslookup foobar.com localhost 5. ls /var/named/sec/foobar.com Actual Results: foobar.com is not copied from the master server unless named is stopped and started again. foobar.com is not mentioned in the log file, the file doesn't appear, and queries for it fail. Expected Results: It should get copied and the server should serve it. Additional info: I do have the following lines in /etc/sysconfig/named: ROOTDIR=/var/named I copied the following stuff into /var/named: /dev/null /dev/random /dev/urandom /etc/named.conf /etc/rndc.conf /etc/rndc.key /var/run The /etc files are symlinked back to the real /etc directory. Everything runs normally, except that rndc reload doesn't load new zones anymore.
Could you check to see if kill -HUP `cat /var/run/named/named.pid` works correctly?
I'm waiting for it to hang again.
Ooops, sorry, wrong bug; disregard that last comment.
I think I found the problem; /var/named/etc was only readable by root. In addition, because /var/named/var/run is owned by named, named.pid is in this directory, not /var/named/var/run/named, so that kill command gave a file not found error. Aren't chroot setups *fun*...