We're running two nameservers on Fedora Core 4. I updated the
secondary nameserver from BIND 9.3.1 to 9.3.5-P1 without a hitch. This
was a standard chroot config (yum -y install bind-chroot) and no yum
update was available so, I had to run configure to ensure that the
update installed into the correct directory tree (I had to do this
twice though - first installed incorrectly).

Once the update was complete - there were no problems on the secondary
nameserver and I ran it for a week to be sure we didn't encounter any
strange issues. That worked with no problem - so I decided to update
the primary nameserver using the same method. Everything seemed fine,
but when I went to restart rndc - I got the following:

[root@ns1 bind-9.3.5-P1]# rndc reload
rndc: neither /usr//etc/rndc.conf nor /usr//etc/rndc.key was found

Now, I noticed the double slash and realized I likely screwed up in
the configuration option by having a slash that shouldn't be there...
so I changed the configuration and did another make and that got rid
of the double slash, but not the problem. So, another sys admin did
the following:
--------------------
1. /usr/etc/rndc.conf recreated. This is the one that rndc uses.
Removed all
others to avoid confusion.

2. Added teh following to named.conf:
key "rndc-key" {
algorithm hmac-md5;
secret "xxxxxxxxxxxxxxxx";
};

3. changed named.conf controls line key to match above:
controls {
inet 127.0.0.1 allow { localhost; } keys { rndc-key; };
};

4. Made sure various diretories under chroot had group write perms.
just
compared to original.
-------------------

My assumption is that somehow the update ran under the wrong
permissions and it broke the chroot jail. Now we have our key and bind
running outside chroot. Can I just start over without losing my
settings?

I'm hoping this makes sense to somebody... I'm pretty new to this and
it's not completely clear to me. Any help is appreciated.