On Sat, Aug 02, 2008 at 10:18:17PM -0700, Mike Bird wrote: > On Sat August 2 2008 18:31:40 firstname.lastname@example.org wrote: > > My network is illustrated here now. > > http://carnot.pathology.ubc.ca/Network.jpg [snip] > > "no connection". > > That routing table shows that dalton knows the route to joule. > From the network diagram we see that dalton also knows the > (connected) route to cantor via eth3 (aka eth1). Assuming no > iptables blocks on any of the three systems, you still need to > ensure: > > 1) That cantor knows the route to joule (via dalton) > 2) That joule knows the route to cantor (via dalton) > > In short, every step along the way needs to know how to route > packets in both directions. something you can do to help track it down is start a ping from cantor to joule and use tcpdump at each of the hop and do a tcpdump -pni <in interface> icmp and host cantor -c 10 and then a tcpdump -pni <out interface> icmp and host cantor -c 10 Trace the packet all the way. on cantor you can start of with ip r g <ip address of joule - note ip r g doesn't take hostname :( > once you know the next hop then you can ssh to there and do the tcpdump thing Alex > > --Mike Bird > > > -- > To UNSUBSCRIBE, email to debian-user-REQUEST@lists.debian.org > with a subject of "unsubscribe". Trouble? Contact email@example.com > > -- Miksch's Law: If a string has one end, then it has another end.
Description: Digital signature