Michael C. Toren on Mon, 29 Oct 2001 02:22:22 -0500 |
> > Send them an angry email, refuse to use their service. Or give in > > and modify your local TCP stack to keep them happy. I know what I'd > > do... > > It's not necessarily them (I haven't checked). It could be any router > in between his machine and the server. One of the neat features of tcptraceroute is the -E flag, which sends out packets with the ECN bits set. So, doing a tcptraceroute with and without the -E flag will pinpoint the hop causing trouble: [mct@quint ~]$ tcptraceroute -q1 www.techrepublic.com Selected device eth0, address 207.8.132.210, port 3527 for outgoing packets Tracing the path to www.techrepublic.com (208.50.157.239) on TCP port 80 (www), 30 hops max 1 jumpgate.townhouse.blue-bell.pa.us (207.106.130.81) 2.089 ms 2 mct-too-l.netaxs.net (207.106.3.253) 4.314 ms [..] 15 64.15.192.36 (64.15.192.36) 74.480 ms 16 64.15.192.122 (64.15.192.122) 75.713 ms 17 * 18 208.50.157.239 (208.50.157.239) [open] 75.896 ms [mct@quint ~]$ tcptraceroute -q1 -E www.techrepublic.com Selected device eth0, address 207.8.132.210, port 3529 for outgoing packets Tracing the path to www.techrepublic.com (208.50.157.239) on TCP port 80 (www), 30 hops max 1 jumpgate.townhouse.blue-bell.pa.us (207.106.130.81) 2.032 ms 2 mct-too-l.netaxs.net (207.106.3.253) 4.286 ms [..] 15 64.15.192.36 (64.15.192.36) 75.795 ms 16 64.15.192.122 (64.15.192.122) 76.286 ms 17 * 18 * 19 * 20 * 21 * -mct
|
|