TeliaSonera routing issue / microsoft.com

Mailman List Mirror (Read Only)
1

Hi -

TeliaSonera has reported routing problem due to prefix filtering leaks / max-prefix triggering, starting around 0616 UTC. I guess some others are seeing this as well. I wonder what was the more exact reason, and why the problem still persists after about 6 hours.

In particular, our customers have noticed this with various sites related to Microsoft, including e.g., VoIP systems failing because contacting some active-x site failed.

Is there a looking glass, traceroute server, or sensible contact info for someone at AS 8075 to check what the routing looks like from that perspective?

(Traceroute works from our upstream, but not from here, which suggests a potential return path issue for e.g., 193.166.0.0/15, pingable address 193.166.255.241).

;; ANSWER SECTION:
www.microsoft.com. 1088 IN CNAME toggle.www.ms.akadns.net.
toggle.www.ms.akadns.net. 57 IN CNAME g.www.ms.akadns.net.
g.www.ms.akadns.net. 57 IN CNAME lb1.www.ms.akadns.net.
lb1.www.ms.akadns.net. 57 IN A 207.46.192.254
lb1.www.ms.akadns.net. 57 IN A 207.46.193.254
lb1.www.ms.akadns.net. 57 IN A 207.46.19.190
lb1.www.ms.akadns.net. 57 IN A 207.46.19.254

  1 s-b4-link.telia.net (213.248.97.93) 0.620 ms 0.498 ms 0.409 ms
  2 s-bb2-link.telia.net (213.248.66.13) 0.762 ms 0.597 ms 0.721 ms
  3 kbn-bb2-pos0-0-0.telia.net (213.248.65.29) 10.231 ms 10.276 ms 10.215 ms
  4 nyk-bb2-link.telia.net (213.248.82.101) 94.660 ms 94.670 ms 94.640 ms
  5 sjo-bb1-link.telia.net (80.91.254.177) 161.549 ms 161.542 ms 161.530 ms

.. stops here.

If it works (from upstream) it continues as follows:

  6 microsoft-110312-sjo-bb (213.248.86.70) 161.421 ms 161.818 ms 157.519 ms
  7 ge-6-3-0-44.sjc-64cb-1a.ntwk.msn.net (207.46.44.98) 161.814 ms 155.456 ms 157.559 ms
  8 ge-1-0-0-0.bay-64c-1a.ntwk.msn.net (207.46.37.158) 156.229 ms 158.292 ms 156.223 ms

2

Pekka Savola wrote:

Hi -

TeliaSonera has reported routing problem due to prefix filtering leaks / max-prefix triggering, starting around 0616 UTC. I guess some others are seeing this as well. I wonder what was the more exact reason, and why the problem still persists after about 6 hours.

This also times with when I lost connectivity to my hosts in the UK ... path is via Above+Telia.

Third hand/Chinese whipsers information:

The UK network(s) are now selectively dropping routes to the affected paths. Telia have been notified.

Regards,

Mat

3

This was the update we'd received.

At the moment we are seeing bit errors on a link between London and New York in a provider�s network, which is significantly affecting traffic on our IP backbone. This issue is escalated to our level 3 manager who is currently in contact with the provider, and we are actively working to reach a resolution. Unfortunately we can not provide an ETR for the issue at this time.

At the same time we are experiencing a fault in London after a planned upgrade. Our IP engineers are currently troubleshooting this issue and we are waiting for further information.