[afnog] mail problem

Thu Apr 20 15:12:56 EAT 2006

On Thu, Apr 20, 2006 at 02:28:39PM +0300, Patrick Okui wrote:
> >If on the Italy side you get them to do a traceroute to 41.220.14.11 this
> > will probably become apparent, as I expect one of the upstream routers will
> > drop the packet with !H or !N (host or network unreachable)
> 
> Not exactly Italy, but from psg I get no joy either (some lines will wrap) ...

Hmm, I think that's just somebody blocking traceroute (or icmp TTL
exceeded), since the actual connectivity is OK:

psg.com:/usr/home/brian> ping -c3 41.220.14.11
PING 41.220.14.11 (41.220.14.11): 56 data bytes
64 bytes from 41.220.14.11: icmp_seq=0 ttl=48 time=702.668 ms
64 bytes from 41.220.14.11: icmp_seq=1 ttl=48 time=740.545 ms
64 bytes from 41.220.14.11: icmp_seq=2 ttl=48 time=704.919 ms

--- 41.220.14.11 ping statistics ---
3 packets transmitted, 3 packets received, 0% packet loss
round-trip min/avg/max/stddev = 702.668/716.044/740.545/17.349 ms
psg.com:/usr/home/brian> telnet 41.220.14.11 25
Trying 41.220.14.11...
Connected to 41.220.14.11.
Escape character is '^]'.
220 mail.one2net.co.ug ESMTP Exim 4.52 Thu, 20 Apr 2006 15:07:41 +0300
quit
221 mail.one2net.co.ug closing connection
Connection closed by foreign host.

Based on the ping results, I would expect the far end to be sending
responses with TTL=64 and therefore it is 16 hops away. This corresponds to
the end of your trace:

> 13  bushnet-111207-kbn-b1.c.telia.net (213.248.97.90)  171.337 ms  171.438 ms  
> 171.233 ms
> 14  217-212-242-2.customer.teliacarrier.com (217.212.242.2)  707.770 ms  
> 701.064 ms  700.154 ms
> 15  217-212-242-13.customer.teliacarrier.com (217.212.242.13)  720.892 ms  
> 713.485 ms  704.651 ms
> 16  * * *
> 17  * * *

Note that the satellite link is between hops 13 and 14. So I think hop 15 is
probably the customer router, and hop 16 is the machine itself, which has
some sort of nasty firewall rules (perhaps just dropping inbound UDP)
preventing the final traceroute hop from appearing.

Come to think of it, you can force traceroute to use ICMP instead of UDP
probes. In which case I get:

psg.com:/usr/home/brian> traceroute -I 41.220.14.11
traceroute to 41.220.14.11 (41.220.14.11), 64 hops max, 60 byte packets
 1  psg2.psg.com (147.28.0.5)  0.428 ms  0.376 ms  0.348 ms
 2  e2.psg1.psg.com (147.28.1.5)  0.723 ms  0.791 ms  0.605 ms
 3  sl-gw11-sea-0-1.sprintlink.net (144.232.9.61)  0.351 ms  0.436 ms  0.353 ms
 4  sl-bb20-sea-9-2.sprintlink.net (144.232.6.125)  139.466 ms  5.073 ms  1.193 ms
 5  sl-bb25-chi-5-0.sprintlink.net (144.232.20.84)  44.835 ms  44.846 ms  44.811 ms
 6  sl-bb23-chi-15-0.sprintlink.net (144.232.26.93)  44.937 ms  44.694 ms  44.689 ms
 7  sl-bb22-chi-13-0.sprintlink.net (144.232.26.49)  44.683 ms  44.967 ms  44.786 ms
 8  sl-st21-chi-11-0.sprintlink.net (144.232.20.21)  45.032 ms  44.982 ms  44.935 ms
 9  144.232.8.98 (144.232.8.98)  46.297 ms  51.641 ms  47.182 ms
10  nyk-bb1-pos0-3-0.telia.net (213.248.80.154)  65.934 ms  66.013 ms *
11  kbn-bb1-pos1-3-0.telia.net (213.248.64.21)  164.724 ms  164.626 ms  164.614 ms
12  kbn-b1-pos1-0.telia.net (213.248.65.10)  291.148 ms  233.103 ms  164.738 ms
13  bushnet-111207-kbn-b1.c.telia.net (213.248.97.90)  172.504 ms  171.422 ms  171.104 ms
14  217-212-242-2.customer.teliacarrier.com (217.212.242.2)  699.784 ms  699.642 ms  702.759 ms
15  217-212-242-13.customer.teliacarrier.com (217.212.242.13)  797.346 ms  727.851 ms  703.910 ms
16  41.220.14.11 (41.220.14.11)  704.276 ms  703.150 ms  707.405 ms
psg.com:/usr/home/brian>

That confirms the theory.

Now we just need someone to do this test from the host which isn't
responding.

Regards,

Brian.