CloudFlare ISTANBUL servers are problematic on docker hub

Hello,

Cloudflare ISTANBUL servers are problematic on docker hub.

Here is the sample CF-RAY from curl output:

480824949caca7ee-IST
480846d97be4b03d-IST
4808496168f5a7b8-IST
48084a7819a7a7ee-IST
48084b4acac6a7b8-IST
48084c22adefb061-IST
48084cb0a8baa7b8-IST
48085a19fe46a806-IST

What kind of errors or unexpected behavior did you experience?

Very slow download from production.Cloudflare.docker.com

I can’t paste the results but 4 out of 5 download attempts fail with some throttling connection for a single image download (21 mb)

ticket on GitHub: https://github.com/docker/hub-feedback/issues/1675#issuecomment-438036421

The ISP in use is TURKCELL SUPERONLINE.

Attempt#1: downloads 21mb image in 2 seconds
Attempt#2: get stuck in %20 21mb image after 40 seconds, waits forever
Attempt#3: get stuck in %17 21mb image after 40 seconds, waits forever
Attempt#4: get stuck in %22 21mb image after 40 seconds, waits forever
Attempt#5: get stuck in %21 21mb image after 40 seconds, waits forever
Attempt#6: downloads 21mb image in 3 seconds

it goes on like this.

I’m not sure whether this is a routing problem of ISP, I collected several trace route logs but I can’t relate with trace is being used in which connection attempt.

All problematic connections seems to be hitting IST servers.

traceroute to 104.18.122.25 (104.18.122.25), 30 hops max, 60 byte packets
1 gateway (10.18.55.1) 1.073 ms 1.036 ms 0.998 ms
2 176.235.179.17 (176.235.179.17) 4.053 ms 4.045 ms 3.997 ms
3 10.38.207.149 (10.38.207.149) 4.303 ms 4.191 ms 5.064 ms
4 10.36.2.81 (10.36.2.81) 11.599 ms 10.40.134.50 (10.40.134.50) 11.591 ms 10.36.2.77 (10.36.2.77) 12.153 ms
5 10.40.137.126 (10.40.137.126) 11.754 ms 12.274 ms 12.313 ms
6 10.38.206.41 (10.38.206.41) 12.546 ms 8.952 ms 8.901 ms
7 10.40.130.106 (10.40.130.106) 3.320 ms 4.550 ms 10.38.211.162 (10.38.211.162) 3.068 ms
8 10.40.130.105 (10.40.130.105) 16.608 ms 10.38.211.153 (10.38.211.153) 15.485 ms 10.40.130.109 (10.40.130.109) 16.548 ms
9 10.36.6.142 (10.36.6.142) 12.970 ms 12.556 ms 12.291 ms
10 10.36.6.38 (10.36.6.38) 28.025 ms 20.021 ms 19.968 ms
11 104.18.122.25 (104.18.122.25) 17.727 ms 17.688 ms 17.679 ms

traceroute to 104.18.122.25 (104.18.122.25), 30 hops max, 60 byte packets
1 gateway (10.18.55.1) 1.019 ms 0.975 ms 0.909 ms
2 176.235.179.17 (176.235.179.17) 3.503 ms 3.466 ms 3.481 ms
3 10.38.207.149 (10.38.207.149) 3.543 ms 4.252 ms 4.307 ms
4 10.36.2.77 (10.36.2.77) 11.379 ms 10.36.2.81 (10.36.2.81) 11.998 ms 10.36.2.77 (10.36.2.77) 12.026 ms
5 10.40.137.126 (10.40.137.126) 11.179 ms 11.802 ms 11.766 ms
6 10.38.206.41 (10.38.206.41) 12.323 ms 9.050 ms 9.006 ms
7 10.38.211.162 (10.38.211.162) 5.254 ms 10.38.211.158 (10.38.211.158) 3.780 ms 3.853 ms
8 10.40.130.109 (10.40.130.109) 11.534 ms 11.536 ms 10.38.211.161 (10.38.211.161) 11.652 ms
9 10.36.6.142 (10.36.6.142) 12.428 ms 11.215 ms 20.732 ms
10 10.36.6.38 (10.36.6.38) 11.736 ms 12.334 ms 12.130 ms
11 104.18.122.25 (104.18.122.25) 14.693 ms 15.424 ms 15.255 ms

traceroute to 104.18.122.25 (104.18.122.25), 30 hops max, 60 byte packets
1 gateway (10.18.55.1) 1.096 ms 0.998 ms 0.987 ms
2 176.235.179.17 (176.235.179.17) 4.746 ms 4.564 ms 4.755 ms
3 10.38.207.149 (10.38.207.149) 4.481 ms 4.533 ms 4.731 ms
4 10.36.2.81 (10.36.2.81) 11.725 ms 11.685 ms 10.40.134.50 (10.40.134.50) 12.189 ms
5 10.40.137.126 (10.40.137.126) 11.651 ms 12.083 ms 12.453 ms
6 10.38.206.41 (10.38.206.41) 12.161 ms 12.290 ms 12.237 ms
7 10.38.211.162 (10.38.211.162) 3.905 ms 10.40.130.110 (10.40.130.110) 5.691 ms 10.40.130.106 (10.40.130.106) 5.605 ms
8 10.38.211.157 (10.38.211.157) 13.186 ms 10.38.211.161 (10.38.211.161) 15.471 ms 10.40.130.109 (10.40.130.109) 14.762 ms
9 10.36.6.142 (10.36.6.142) 12.845 ms 9.793 ms 9.709 ms
10 10.36.6.38 (10.36.6.38) 11.302 ms 11.615 ms 11.481 ms
11 104.18.122.25 (104.18.122.25) 14.474 ms 18.746 ms 18.689 ms

There are several trace routes, all end up with the same router and ip but intermediate routes changes. I can’t relate why intermediate routes affect the connection throttling.

This topic was automatically closed after 14 days. New replies are no longer allowed.