using ndc to reload/reconfig bind that is chroot'ed

[Phil N]

Ralf -

Late last night I found the problem.  The parent directory of "/etc" (from the
chroot perspective) needed to be owned by the user "named".  For some reason
named would start just fine but if I sent it a reload signal, it couldn't read
the /chroot/named/etc/named.conf file when the permissions on the /chroot/named
directory were not set as indicated below.  I did have the directory owned by
root.root with 0744 permission.....

This behavior, to me, was interesting since in the INSTALL file notes indicate
that:

      "-t" followed by a directory path on the "named" command line will
        cause the server to chroot() to that directory before it starts
        loading the configuration file.

Based on that (from my limited experience perspective :), I don't think that the
server should have started at all.

Phil

******
options {
        directory "/etc/namedb";
        pid-file "/var/log/named.pid";
};
******
[root at dns1 /chroot]# pwd
/chroot
[root at dns1 /chroot]# ls -lt
total 8
drwxrwx---   2 root     root         4096 Jun 12 16:43 postfix
drwxrwx---   8 named    named        4096 May 18 09:22 named
[root at dns1 /chroot]#
*****


> Strange. Are the permissions ok (readable for user named, group named)?