Slave servers not auto-updating? (BIND 9.2.1)

[Mark Damrose]

"Don Werve" <devnull at examen.com> wrote in message
news:slrnasns18.fme.donw at gir.agentsix.net...
> I've got an odd little problem, which is more of an annoyance than a
> mission-critical problem.  I've got two slaves sitting off a master
> server (all running the same version of BIND), and when I make an update
> to one of the zonefiles on the Master and 'kill -HUP' it, only one of
> the slaves updates its own cache of the updated zone.
>
> Yes, I'm updating the serial numbers, and all three servers sync off the
> same NTP server.  Two are Linux boxen, the third is a Solaris 8 machine,
> but all three are running a hand-compiled copy of BIND (chrooted for
> security).
>
> Nothing odd shows up in the logs on the Solaris machine (the one that
> doesn't update), and if I restart the daemon on that box, it then grabs
> the update without a hitch.

If you let it wait until the REFRESH timer, does it update?  Is it listed
with an NS record?  If not, do you have an also-notify { ip of Solaris; };
configured?

>
> Any ideas why?  I've RTFM and didn't see anything helpful, but I am only
> a mere mortal and might have missed something, so if there are any
> pointers to relevant docs, I'd like them.
>
> Thanks-in-advance!
>
> --
> Don Werve (Unix Sys Admin) | Email: donw AT examen dot com
>
> "Whatever you think you can do or believe you can do, begin it.  Action
> has magic, grace and power in it." -- Goethe