I have a Redhat system that runs its own dns which I maintain with rpm's. At some points I loose the ability to resolve dns but still have comms with my ISP, and the next time that happens I'll change my cisco pix log level to "warn" to see what's being rejected. In any event, as I was searching for some problem with my Redhat system, I uninstalled Bind9 and installed the latest version from source to see if the problem would go away. It did not, so I uninstalled Bind9 again and reinstalled the latest Bind rpm from Redhat, since it's easier to manage system updates in this way. Since that time, the named daemon on my RHES system will not start on boot, although when I type #service named start in a terminal window, Bind starts without error. I haven't found any dmesg or /var/log/messages output that points to the problem, and I'm not experienced enough to look at the boot script in /etc/init.d/named to figure things out.