Description of problem: I upgraded to rawhide today, this installed: caching-nameserver-9.3.2-37.fc6 Immediately, NetworkManager's nameserver broke, and restarting either named or NM did not fix it. I eventually just edited /etc/resolv.conf so I could get things done.
I have observed this behaviour as well. But rather than editing /etc/resolv.conf, I have found /sbin/service named stop works as well. When stopping, real values are written to resolv.conf automatically instead of the 127.0.0.1 that NetworkManager installs.
If you don't restart named manually, which networks aren't available? Could you attach bind configuration?
At this point, it's been reinstalled twice and NM has been downgraded. Not sure I can reproduce.
How many network devices do you have? Do you use any VPN?
Wired and wireless, although only one is ever active at a time. Yes, I use VPN, (NM-vpnc) and I believe it was up at the time.
okay, and do you have still this problem or it appeared only after the upgrade? It's probably a variant of #203291 or #196962
Haven't seen it since then, no.