Date: Tue, 01 Aug 2006 13:29:04 +0200 From: Edda Hochstrate <eh@netuse.de> To: Doug Barton <dougb@FreeBSD.org> Cc: freebsd-isp@freebsd.org Subject: Re: [SPAM] (4.4/3.0) Re: Bind9: rndc reload doesn't work for slave servers Message-ID: <44CF3B00.2090106@netuse.de> In-Reply-To: <44CE5875.9050509@FreeBSD.org> References: <44CE2043.1040308@netuse.de> <44CE5875.9050509@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Doug Barton wrote: > Edda Hochstrate wrote: >> >>In my opinion afer an "rndc reload" a slave >>should query the master for the soa records of all >>zones stated in named.conf. > > > Well, unfortunately for you the BIND developers do not agree with you. I > could argue that there are good reasons not to do what you're suggesting, > but I suspect that is not a useful discussion. Thank you, that clears the situation. I was very surprised about this behaviour. > > Now that I've answered your question, I'd like to suggest that you perhaps > reconsider this goal? If your master and slave name servers are properly > configured, you shouldn't have to do this at all. The master should be > sending out notifys for domains when they are updated, which your slaves can > then act on in near real time. What conditions are you seeing that lead you > to believe that refreshing all your zones at once is necessary? We have a lot of customers with master servers for their own domains, not every server is bind 9 or even bind. The notify mechanism doesn't work well on all these servers. With bind 8 we simply send a kill -HUP to our slave and it starts reloading all updated zones. With bind 9 we observe that our slaves get some updates only after the refresh time is expired. Thank you again, Edda
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?44CF3B00.2090106>