Re: [netatalk-admins] Odd nbprgstr problems with 1.4b2/FreeBSD 2.2.5


Subject: Re: [netatalk-admins] Odd nbprgstr problems with 1.4b2/FreeBSD 2.2.5
From: Alistair Riddell (alistair@watsons.edin.sch.uk)
Date: Fri Mar 13 1998 - 13:07:30 EST


Are you sure your atalkd.conf file matches the network? If you move from a
network with routes to one without withou removing or changing atalkd.conf
then atalkd might get confused.

On Fri, 13 Mar 1998, Peter Hjelt wrote:

> Got a itchy problem with netatalk 1.4b2 straight out of the ports collection
> of FreeBSD 2.2.5, using the stock appletalk drivers from the 2.2.5 kernel.
>
> I installed the server, and brought it up on my production-network for
> testing, a network full of busy macs, afp servers, routers, zones, whatnot
> .. and it worked flawlessly from the start. Very smooth, in fact.
>
> Now, once I move it to a more quiet network with just one or two macs,
> possibly an old afp server .. or a network with just one idle mac (7.5.3),
> starting out with a nulled atalkd.conf, I get the following error:
>
> # nbprgstr -p 4 foo:Workstation
> npb_rgstr: Operation timed out
> Can't register foo:Workstation@*
>
> nbprgstr broadcasts the name (because the mac worksations immediately see
> it in their choosers), but then nbprgstr is expecting a reply it is not
> getting for one reason or another.
>
> It's timing out in the select() call at line 109 of nbp_rgstr.c when
> waiting for appletalk datagrams as a response to the broadcast annoucing
> the name ..
>
> I'm not sure if there are any packets directed to the server as I don't
> have a sniffer able to analyze appletalk .. and the kernel driver for
> netatalk/appletalk for FreeBSD has no debug/trace options. Still, I don't
> see any reason the sock macs on the network would not respond as supposed to.
>
> I increased the timeout to 20 seconds just to make sure, but it obviously
> never receives any packets on the appletalk datagramsocket listening for
> ATADDR_ANYNODE.
>
> The only odd part about this is that the server works (!) when moved back
> to my busy network, if it didn't I would be suspecting the kernel drivers,
> NIC, and whatnot. The server uses a 3Com 900 XL NIC (vx device).
>
> What am I overlooking? What response is it expecting?
>

--
Alistair Riddell - BOFH
IT Support Department, George Watson's College, Edinburgh
Tel: +44 131 447 7931 Ext 176       Fax: +44 131 452 8594



This archive was generated by hypermail 2b28 : Sat Dec 18 1999 - 16:31:38 EST