Wouter Klouwen <dublet%acm.org@localhost> wrote: > The rc script stores the pid of the named process in a file in /var > somewhere, and uses that to control the named process. This assumes > there is only one. Therefore if you launch named without the rc > script, doing a 'sh /etc/rc.d/named stop' won't do anything, and a > 'start' will just start a new instance. So you will indeed have to > use kill to terminate the "unmanaged" process, otherwise the > situation will just continue, no matter how many starts, stops or > restarts you issue. > > Somehow named was either started manually or /var/run/named.pid got > lost. Well, I know that, but I'm pretty sure I never started named manually, that's what confused me. I have, however, /etc/rc.d/named stop in my /etc/ppp/ip-down and /etc/rc.d/named stop in my /etc/ppp/ip-up, so it seems something went wrong there. Anyway, the reason I do that is that named complains when the connections is lost and adds a message to syslog about each message. This is more an ugly workaround than anything else. So I think I should find out why it does that and fix that. -- Jonathan
Attachment:
signature.asc
Description: PGP signature