Date: Fri, 21 May 2004 13:15:26 +0100 From: Matthew Seaman <m.seaman@infracaninophile.co.uk> To: Nuno =?iso-8859-1?Q?C=E9sar?= Pires <NCPires@isq.pt> Cc: "'freebsd-questions@freebsd.org'" <freebsd-questions@freebsd.org> Subject: Re: Help to start BIND on boot Message-ID: <20040521121526.GD31695@happy-idiot-talk.infracaninophile.co.uk> In-Reply-To: <6A7539CE32D1CC42AD0B863337EEFCFB020B20F6@isq004> References: <6A7539CE32D1CC42AD0B863337EEFCFB020B20F6@isq004>
next in thread | previous in thread | raw e-mail | index | archive | help
--qGV0fN9tzfkG3CxV
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable
On Fri, May 21, 2004 at 12:07:25PM +0100, Nuno C=E9sar Pires wrote:
> I'm setting up a BIND/DNS server (recursive) and I'm facing a problem whe=
n I
> try to start the "named" automatically on boot (named_enable=3D"YES" in
> rc.conf):
>=20
> First I noticed that the boot process takes a very long time in the
> "Starting Standard Daemons" phase.
>=20
> After boot I have looked to the named messages and apparently it was
> everything ok i.e.:
>=20
> =20
>=20
> First message: "starting (etc/namedb/named.conf). named
> 8.3.6-REL ....."
>=20
> Second message: "limit files set to fdlimit (1024)"
>=20
> Third message: "Ready to answer queries."
>=20
> =20
>=20
> But the problem is the output of "ndc status": (server is initializing
> itself) and then nothing happen.
>=20
> =20
>=20
> Starting or restarting the named manually works just fine, the "ndc statu=
s"
> output is "server is up and running" and the queries answers are as
> expected.
>=20
> The result after comment the kern_securelevel=3D"2" and
> kern_securelevel_enable=3D"YES" lines in the rc.conf and reboot was the s=
ame
> as described above.
>=20
> =20
>=20
> The ROOT SERVERS file was updated and there is a permanent network
> connection
>=20
> =20
>=20
> Releases:
>=20
> FreeBSD 4.9
>=20
> BIND 8.3.6
I usually see this effect with things like sshd(8), but it could
affect BIND as well. I wonder if named(8) is blocking trying to read
/dev/random to obtain a quantity of random data. If the system does
not have sufficient suitable random data available, it will wait until
it has acquired enough before replying. Sources of randomness are
things like timing the gaps between key presses or between the arrival
of network packets -- either of which may not be very effective around
reboot time.
Check your setting for 'rand_irqs' in /etc/rc.conf -- you need to set
it to a list of IRQs that fire quite frequently and that have timings
that can be used to harvest randomness from. To get a list of
suitable IRQs use:
% vmstat -i
So for instance on my system that returns:
interrupt total rate
acpi0 irq9 1 0
pcm0 irq10 39644 0
mux irq11 12139824 77
mux irq15 854820 5
atkbd0 irq1 49505 0
psm0 irq12 389549 2
sio1 irq3 81928 0
clk irq0 157097139 1000
rtc irq8 20105805 128
Total 190758215 1214
Choose the IRQs that fire most often, but not the clk (clock) or rtc
(real time clock) IRQs -- as those fire at regular intervals. In this
case good choices are irq1 (atkbd -- the keyboard), irq11 and irq15
(mux -- the TCP multiplexor (ie network traffic)), irq12 (psm -- the
mouse).
Cheers,
Matthew
--=20
Dr Matthew J Seaman MA, D.Phil. 26 The Paddocks
Savill Way
PGP: http://www.infracaninophile.co.uk/pgpkey Marlow
Tel: +44 1628 476614 Bucks., SL7 1TH UK
--qGV0fN9tzfkG3CxV
Content-Type: application/pgp-signature
Content-Disposition: inline
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (FreeBSD)
iD8DBQFArfLeiD657aJF7eIRAnylAKCY6cVv8WsY9yH+I56YvsrbUNdbxACff781
b8BJXUsTh8tNFSWSJ9R46fs=
=iMMP
-----END PGP SIGNATURE-----
--qGV0fN9tzfkG3CxV--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20040521121526.GD31695>