Re: named support now requires passing --with-named



On Fri, 2005-01-14 at 15:40 -0500, Daniel Gryniewicz wrote:

> Will it flush the cache when the subnet changes?

In short, not yet.  It's definitely a bug.  I hit this when doing VPN
stuff, because internal *.redhat.com addresses do not resolve when I'm
at home and not on the VPN, so the negative existence gets cached, and
when I get on the VPN those hostnames appear to not exist until I
remember to flush the cache.

Unfortunately it's hard to fix, because flushing the cache requires
rndc, and doing rndc requires listening on a TCP port.  Argh! So I'd
have to try allocating a random TCP port, in order to avoid conflicting
with the system rndc.  Ugh.  rndc would definitely be better done using
D-BUS.

I kind of wish more Unix daemons were written as shared libraries from
the start.

Anyways, I could probably just enable rndc for now, and assume that the
system isn't also a BIND server.  Probably a reasonable assumption for
the current NetworkManager userbase, but longer term we need to fix it.



Attachment: signature.asc
Description: This is a digitally signed message part



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]