Skip to content

Dallas location suddenly routing through Tokyo, anyone else seeing this?

VPS Hosting by kat_fold 27 replies 3.2K views
#21

kat_fold said:
Mexico City is unlikely to show the same path unless your provider peers through Seattle.

I said it sounds like a misread because 180ms is exactly what I would expect for a real trans-pacific path. The surprising part is that it comes back to St. Louis after Tokyo, not the number itself. 180ms to Tokyo is normal. 180ms to St. Louis via Tokyo is the problem.

I see the same from my location. Not a misread.

7 #22

adadude said:
Idempotent results

This word does not mean what you think it means. Idempotent means applying the operation twice has the same effect as once. You mean consistent or reproducible results.

The routing is consistently wrong. Your vocabulary is inconsistently wrong.

#23

Head13 said:
This word does not mean what you think it means.

In the context of a cron job, idempotent means the script can run multiple times without side effects, which it does. The results are also consistent. Both words apply. I will not apologize for compound accuracy.

The routing remains via Tokyo as of 14:00 UTC. No change.

#24

Following this with interest. My HostHatch Stockholm to Contabo St. Louis is normal at 42ms. No anomalies on European paths.

Has anyone contacted NTT directly? They are the common jp hop in most of these traces.

#25

JohnCreek said:
Has anyone contacted NTT directly?

I opened a ticket with Contabo, they said upstream issue. I do not have a direct NTT contact. If this is a peering dispute as they said, NTT will not talk to end users anyway.

My workaround through HostHatch Chicago is still holding. 78ms Madrid to Chicago, 15ms Chicago to St. Louis via the tunnel. Total path is longer geographically but shorter in practice.

swimming upstream since 2019 🐟
#26

I am seeing recovery. traceroute from my location now shows denver -> chicago -> st. louis, no Tokyo. Latency back to 32ms as of 15 minutes ago.

Can anyone else confirm?

#27

StevenNmp said:
Can anyone else confirm?

Confirmed from Cleveland. Path normalized at approximately 16:30 UTC. My cron'd mtr shows the jp hop disappeared between 16:25 and 16:30.

Total duration roughly 12 hours. Someone finally fixed the leak or withdrew the bad announcement.

#28

Marking this as resolved based on multiple confirmations. 12-hour routing anomaly, cause upstream peering issue per provider tickets, no action required by end users beyond waiting or tunneling around it.

If it happens again, open a new thread. Locking this one to prevent necromancy.

Nadia59 said:
Is there a website that does this?

For future reference: ping.pe, mtr.sh, or your own VPS with mtr installed.

~be kind or be gone~

Post a reply

You need an account to reply. Log in or register to join the conversation.

Post reply Preview Save draft