I was recently made aware of a problem with traceroute6 from the iputils-tracepath package, and I've only been able to duplicate the problem when running kernel 2.4.19. But, that's not to say that *all* 2.4.19 systems are effected; the laptop I'm using now doesn't exhibit the problem under 2.4.19, but my workstation does. The symtoms of the problem are traceroute6 outputs that look like: ant:~$ traceroute6 -n www.linux-ipv6.org traceroute to moon.linux-ipv6.org (2001:200:0:1003:200:f8ff:fe1e:2d69) from 2002:d03b:fa68:dead:2a0:ccff:fe57:a34e, 30 hops max, 16 byte packets 1 98fc:ffbf:4e51:1240:e0a0:408:70a2:408 0.258 ms 0.163 ms 0.159 ms 2 98fc:ffbf:4e51:1240:e0a0:408:70a2:408 479.163 ms 472.151 ms 483.137 ms 3 98fc:ffbf:4e51:1240:e0a0:408:70a2:408 449.042 ms 461.665 ms 469.646 ms 4 98fc:ffbf:4e51:1240:e0a0:408:70a2:408 309.023 ms 315.5 ms 314.552 ms 5 * 2002:d03b:fa68:dead::1 409.508 ms 422.063 ms 6 2002:d03b:fa68:dead::1 411.045 ms 418.124 ms 417.661 ms 7 2002:d03b:fa68:dead::1 383.046 ms 383.544 ms 396.091 ms 8 2002:d03b:fa68:dead::1 388.026 ms 409.155 ms 386.63 ms 9 2002:d03b:fa68:dead::1 403.959 ms 422.665 ms 430.162 ms 10 2002:d03b:fa68:dead::1 384.55 ms 396.111 ms 379.592 ms I can watch network traffic with tcpdump and see that the right bits are going over the network, but completely incorrect addresses are printed. 2002:d03b:fa68:dead::1 is the IPv6 gateway, and should be the first hop. 98fc:ffbf:4e51:1240:e0a0:408:70a2:408 is obviously bogus. tracepath6 does not show this problem, and if I reboot the system to 2.4.17 traceroute6 works fine. Has anybody seen anything like this before? noah -- Noah Meyerhans Computer Resource Services, MIT Laboratory for Computer Science
Attachment:
pgp1bKWSSj6QT.pgp
Description: PGP signature