Traffic levels are back to normal. On my way to pick up my kid from school, Savvis called to verify the issue had been resolved. On Wed, Apr 24, 2013 at 2:27 PM, Jeremy Chadwick <jdc@koitsu.org> wrote:
Todd,
http://www.nanog.org/meetings/nanog47/presentations/Sunday/RAS_Traceroute_N4...
I'll be more cognizant of these things, even when hair mode = fire.
1. Can you provide source and destination IPs?
nat'd src 208.89.136.20 to dst 98.139.183.24
2. Do you have a return path traceroute? This is incredibly important.
I was in too much of a hurry. You should be able to traceroute back to 64.14.201.31 or 208.89.138.21.
3. Can you use a utility like mtr instead, from both endspoints?
Next time I'll go out and back from our rackspace machine in HK. Obviously I would have to use something different as the remote since my trace was to yahoo.com and I am not a yahoo employee. I just picked something different from my co-worker (he used google) to make sure that in saying "the Internet is down", I wasn't just some idiot if it was just "google is down".
4. Is there some reason your below traceroute is in some format that looks almost hand-written and lacks any indication of packet loss (e.g. no "*" indicators)? It's not that I don't believe you, it's that the evidence needs to be presented.
Agreed. I used tracepath (CentOS 6.x) instead of mtr or traceroute. Again, too much of a hurry.
If you need a destination to reach (that hopefully utilises the network path you're seeing problems with), you can use my box: 206.125.172.42. I do not filter ICMP. If you provide an answer to #1 I can provide return-path results.
Thanks for the offer, but no need. The master ticket has been resolved, I'll know more technical details in 5 days or less.
But in general, testing against a site where you don't have return-path access/visibility makes troubleshooting very difficult. See the PDF I linked. :-)
Agreed. Too much going on at once and didn't get all the info out I should have. Thanks for the reply Jeremy, it's much appreciated. ...Todd -- The total budget at all receivers for solving senders' problems is $0. If you want them to accept your mail and manage it the way you want, send it the way the spec says to. --John Levine