I stupidly screwed up my pf.conf, as a result "ntpd -s" which is invoked in /etc/rc (as a result of my /etc/rc.conf.local) could not resolve the names of the time servers.
ntpd hangs and cannot be interrupted. The only way to continue is to do a hardware reset. I realize that it was my mistake but most of the other processes started by /etc/rc can be interrupted. I really do not like doing hardware resets. As an aside, it was on my firewall. My firewall makes use of my dns which is on the inside of my network, but during the booting process the pf.conf cannot refer to a dns name that are define on the outside. The startup pf.conf built into /etc/rc allows dns requests originating from machine being booted only. For an external name, my dns has to pass a request though the firewall to the outside dns server. That cannot be done until the system is fully booted, The solutions are: 1) don't use any external dns names in your pf.conf 2) have three stage bootstrap of pf, First stage, the code in /etc/rc Second stage, /etc/rc.conf.local has pf.conf to allow the inside nameservers to pass though Third stage, /etc/rc.local has code to set the final pf.conf file 3) put all external dns names in tables that have their contents defined in /etc/rc.local (this is the one I currently use) 4) Modify /etc/rc to allow pass through of dns requests. The ones that should be allowed pass though would be the nameservers defined in /etc/resolv.conf (this is what I would like)
