From: Tomi Manninen (tqgr@mx.dy.fi)
Date: Mon May 01 2000 - 18:46:27 EEST
On Mon, 1 May 2000, Kjell Jarl wrote:
> I do the same, have both netrom and port call the same, and letting node
> respond to ax.25 user connects to the same SSID.
> It seems to work, and makes it obviuos for users what to connect to.
> Maybe some one could elaborate on why not to do it?
If you let an application like LinuxNode answer to L2 requests with the
same callsign that is used for internode NET/ROM traffic you can have
problems like this:
Your neighbor node wants to send NET/ROM traffic to you. It needs to open
a L2 connection to you and it uses the callsign shown on your NODES
broadcasts (== first nrports callsign == HWaddr for nr0). But now you have
configured ax25d to listen for that callsign and to lauch "node". Linux
has no way of knowing so it does what it is told and node is lauched. Node
then sends it's welcome text and the trouble begins...
What happens next depends on what is at the other end of the connection
(the neighbor) and it ranges from a minor inconvenience through lots of L2
disconnetions and lost L3 frames to an avalanche of "invalid command"
messages flooding the band.
Ever wondered why node software like BPQ and TheNet never send a welcome
message...? :-) (And no, just dropping the welcome text from LinuxNode is
not a solution though I have been thinking of adding that as an option
to node.)
The same thing applies to AX.25 interface callsigns (those in axports) if
you expect to have incoming TCP/IP traffic using Virtual Circuit mode.
-- Tomi Manninen Internet: bkjm@high-tower.com OH2BNS AX.25: terhi.victor@logonet.com KP20ME04 Amprnet: zoqbhm.ypkrbjro@gfos.hr
This archive was generated by hypermail 2b29 : Mon May 01 2000 - 18:52:43 EEST