[CUWiN-Dev] Not routing to a connected user... missing something?

Rob Simmons robsimmons at gmail.com
Wed Apr 5 16:19:58 CDT 2006


This is a (probably not very good) slice of the network topology that I'm
dealing with, but I don't know if that element's too important. I currently
have someone connected to a node (laptop, 10.0.106.141) that can connect two
nodes which both connect to the tower in some number of hops, often just 0. 

West Tower <-> (0 or more links) <-> 10.0.241.179 <------> 10.0.106.141 
                          \--------> 10.0.241.220 <-/

The sip0 interface on 10.0.106.141 is 10.20.21.254, and the IP address of
the user's machine is 10.20.21.252. I can ping 10.20.21.252 from the node
only - they seemed to have some success pinging from their computer out to
10.0.241.220 and other nearby nodes, but they definitely could not get out
to the network, even though when I had left the node the night before it
worked fine.

It doesn't seem to be their computer's fault, since I can ping it locally,
and she claimed to even be receiving pings back from nearby nodes when I
talked it over with her on the phone.

When I try to ping 10.20.21.254 (the local interface of the node), it works
from pretty much anywhere. In every routing table I checked the next-hop
entries for 10.20.21.254 and 10.20.21/24 were the same.

I confirmed that I could ping other computers remotely from elsewhere around
the network, but when I try to ping 10.20.21.252 from anywhere besides the
node it is connected to, I get lots and lots and lots of silence, but
occasionally some messages in the style of:

root at 10.0.241.220 # ping -R 10.20.21.252
PING 10.20.21.252 (10.20.21.252): 56 data bytes
ping: sendto: No buffer space available
ping: sendto: No buffer space available
ping: sendto: No buffer space available
ping: sendto: No buffer space available
ping: sendto: No buffer space available
76 bytes from 10.0.233.19: Time To Live exceeded for icmp_seq=50
ping: sendto: No buffer space available
76 bytes from 10.0.233.19: Time To Live exceeded for icmp_seq=51
ping: sendto: No buffer space available
ping: sendto: No buffer space available
ping: sendto: No buffer space available
76 bytes from 10.0.241.179: Time To Live exceeded for icmp_seq=455

Where 10.0.233.19 is another random node on the network that wouldn't seem
to have much significance to anything.

A bit of haphazard or slightly systematic rebooting hasn't seemed to change
anything with this. 

Simultaneously with this the ping times for a lot of the network are up - in
some places to about six seconds in a couple of places that used to have
very low ping times, but rebooting the west tower gateway appeared to help
that a bit.

Does anybody have a better idea of what I should or could be looking for
here?

 - Rob



More information about the CU-Wireless-Dev mailing list