Has anyone ever seen this?

[Mathias Körber]

> On Thu, 9 Nov 2000, O'Neill, Mike wrote:
>=20
> >=20
> > I have a secondary name server running Redhat (release 5.0 =
(Hurricane))
> > I'm running =3D>BIND "8.1.1"  =20
>=20
> Yummie !!! An ancient version of Bind... Something to hack tonight ;PP =


I agree. He should update to 8.2.2-P7 (out today at www.isc.org)

>=20
> > On the console the following message continues to scroll:
> >=20
> > error get_socket:
> > Bind 49 address already in use
> > Init: id "ta" respawning too fast: disabled for 5 minutes
> >=20
> > Any ideas on how I can get this corrected? =20
>=20
> Kick named OUT of /etc/inittab. Daemons like named don't belong there, =
it
> is VERY, VERY, VERY, VERY UGLY.

1. It does not look to me that that is a BIND message. It looks more to =
me
that it refers to the system-call bind(2) and could be any other program
in inittab.

2. Why is running bind from /etc/inittab bad? It has a -f option for =
just
that purpose. A good way of making sure there always is a named up and =
running.
(the other one is putting "ndc start" into a crontab executed every 10 =
or so
minutes).