Bugzilla – Full Text Bug Listing
| Summary: | Really slow network connection | ||
|---|---|---|---|
| Product: | Community | Reporter: | Chris Frost <eclipse> |
| Component: | Servers | Assignee: | Eclipse Webmaster <webmaster> |
| Status: | RESOLVED FIXED | QA Contact: | |
| Severity: | normal | ||
| Priority: | P3 | CC: | hsiliev |
| Version: | unspecified | ||
| Target Milestone: | --- | ||
| Hardware: | PC | ||
| OS: | Mac OS X - Carbon (unsup.) | ||
| Whiteboard: | |||
|
Description
Chris Frost
Can you help me understand what "pretty quick" and "slower" mean, in terms of numbers? Also, to what host are you uploading? Also, I'd need your remote IP address to see if there's anything going on across the world. Feel free to email it to webmaster@eclipse.org Looks like we're picking up some major lag on the backbone between Toronto and the UK: 8 bb1.tor.primus.ca (195.69.145.154) 11.602 ms 11.223 ms 10.470 ms 9 ams-ix-1.enta.net (195.69.145.82) 177.651 ms 177.412 ms 178.869 ms How did you detect this, a colleague in Bulgaria is having the problem. Guess it could be the same pipes causing it. A tcp traceroute from build.eclipse.org to your location. If you send me their IP in Bulgaria I can have a look. However, it's pretty far down the route, and there's probably not much I, or our ISP can do about it. Hi, I get a log of timeouts trying to clone git repos. My IP is 91.92.153.114. From my side: PING git.eclipse.org (206.191.52.44) 56(84) bytes of data. 64 bytes from 206.191.52.44: icmp_req=1 ttl=111 time=161 ms 64 bytes from 206.191.52.44: icmp_req=2 ttl=111 time=161 ms 64 bytes from 206.191.52.44: icmp_req=3 ttl=111 time=161 ms traceroute to git.eclipse.org (206.191.52.44), 30 hops max, 60 byte packets 1 router (192.168.1.1) 0.742 ms 1.001 ms 1.335 ms 2 91.92.153.113 (91.92.153.113) 4.921 ms 4.914 ms 5.089 ms 3 rtr-int.cores.bg (85.187.217.193) 6.923 ms 7.097 ms 7.089 ms 4 95-43-228-233.btc-net.bg (95.43.228.233) 7.452 ms 7.690 ms 7.686 ms 5 193.231.106.37 (193.231.106.37) 12.205 ms 12.379 ms 12.981 ms 6 10.0.241.133 (10.0.241.133) 58.040 ms 54.701 ms 54.935 ms 7 10.0.244.165 (10.0.244.165) 56.587 ms 56.583 ms 56.627 ms 8 10.0.200.14 (10.0.200.14) 58.340 ms 58.336 ms 58.562 ms 9 10.0.200.38 (10.0.200.38) 56.787 ms 54.438 ms 56.162 ms 10 rc2ec.il.shawcable.net (195.69.144.148) 156.567 ms 158.322 ms 158.116 ms 11 66.163.65.1 (66.163.65.1) 168.157 ms 162.658 ms 163.622 ms 12 rc2sh-pos3-0.mt.shawcable.net (66.163.76.146) 160.273 ms 159.245 ms 159.822 ms 13 ra1fs-tge2-3.mt.shawcable.net (66.163.66.122) 160.301 ms 160.788 ms 161.005 ms 14 rx0sh-mountain-cable.mt.bigpipeinc.com (66.244.223.238) 155.685 ms 155.675 ms 155.817 ms 15 10.201.120.133 (10.201.120.133) 165.084 ms 165.052 ms 162.972 ms 16 10.201.120.77 (10.201.120.77) 165.011 ms 164.709 ms 164.436 ms 17 10.201.122.194 (10.201.122.194) 165.405 ms 163.798 ms 168.639 ms 18 206.191.0.111 (206.191.0.111) 164.608 ms 164.812 ms 165.133 ms Regards, Hristo Iliev All is well up to Romania: > 5 193.231.106.37 (193.231.106.37) 12.205 ms 12.379 ms 12.981 ms This next hop somewhere between Romania and Amsterdam is adding 40+ ms. I can't trace it since it's a private IP. > 6 10.0.241.133 (10.0.241.133) 58.040 ms 54.701 ms 54.935 ms > 7 10.0.244.165 (10.0.244.165) 56.587 ms 56.583 ms 56.627 ms > 8 10.0.200.14 (10.0.200.14) 58.340 ms 58.336 ms 58.562 ms > 9 10.0.200.38 (10.0.200.38) 56.787 ms 54.438 ms 56.162 ms Amsterdam -- ouch. +100ms > 10 rc2ec.il.shawcable.net (195.69.144.148) 156.567 ms 158.322 ms 158.116 ms The hop from Amsterdam to Canada is only adding 10ms, which is nice to see. > 11 66.163.65.1 (66.163.65.1) 168.157 ms 162.658 ms 163.622 ms The rest of the pipes in Canada aren't adding any significant lag. > 18 206.191.0.111 (206.191.0.111) 164.608 ms 164.812 ms 165.133 ms From what I can tell, some of the European/atlantic pipes are pretty busy. Things are much worse today: 9 bb1.tor.primus.ca (195.69.145.154) 18.772 ms 17.752 ms 18.275 ms 10 ams-ix-1.enta.net (195.69.145.82) 184.808 ms 186.518 ms 186.507 ms I've opened a ticket with our ISP (primus.ca) to see if they can change the routing. That lag is ridiculous. > Amsterdam -- ouch. +100ms
> > 10 rc2ec.il.shawcable.net (195.69.144.148) 156.567 ms 158.322 ms 158.116 ms
So the common link between Chris' lag and Hristo's lag is that it happens in Amsterdam -- albeit from two different routes.
I've forwarded this finding to our ISP. They should be able to route differently from Toronto.
Hi, Thank you for keeping an eye on this. It is appreciated. Chris. (In reply to comment #8) > > Amsterdam -- ouch. +100ms > > > 10 rc2ec.il.shawcable.net (195.69.144.148) 156.567 ms 158.322 ms 158.116 ms Our ISP confirms a new gateway has been selected. Looks like hop #10 is now a lot faster. 10 ams-ix-1.enta.net (195.69.145.82) 98.411 ms 97.991 ms 99.117 ms This may not be earth shattering, but the extra 60ms per packet was contributing to lots of lag. I'll keep an eye on it for a few days to make sure it's not just a time-of-day thing. If you can report back your observations that would be good. Hi, At the same time as you sent comment 8 I had a local build running and it sped through with no problems. It makes me think the network lag is a different issue unless the ISP had already switched the route. Either way, things are looking much better now, thank you. Chris. I was able to clone all Virgo repos without problems. Looks better now. traceroute to git.eclipse.org (206.191.52.44), 30 hops max, 60 byte packets 1 router (192.168.1.1) 0.896 ms 1.128 ms 1.382 ms 2 91.92.153.113 (91.92.153.113) 5.374 ms 5.367 ms 5.558 ms 3 rtr-int.cores.bg (85.187.217.193) 6.811 ms 6.982 ms 6.975 ms 4 95-43-228-233.btc-net.bg (95.43.228.233) 7.591 ms 7.611 ms 7.572 ms 5 193.231.106.37 (193.231.106.37) 14.524 ms 14.324 ms 14.503 ms 6 10.0.241.133 (10.0.241.133) 57.936 ms 53.684 ms 54.097 ms 7 10.0.244.165 (10.0.244.165) 56.121 ms 56.113 ms 56.349 ms 8 10.0.200.14 (10.0.200.14) 58.331 ms 58.598 ms 58.590 ms 9 10.0.200.38 (10.0.200.38) 54.451 ms 52.851 ms 53.959 ms 10 bb1.tor.primus.ca (195.69.145.154) 142.912 ms 142.907 ms 143.137 ms 11 10.201.120.129 (10.201.120.129) 152.201 ms 151.491 ms 154.223 ms 12 10.201.120.77 (10.201.120.77) 152.547 ms 154.434 ms 156.885 ms 13 10.201.122.194 (10.201.122.194) 153.002 ms 154.396 ms 157.650 ms 14 206.191.0.111 (206.191.0.111) 153.659 ms 153.186 ms 150.437 ms Chris, Hristo, are you still affected by this, or have the routing and peering changes helped? Hi, It's been OK for me recently so I guess so. Ok, let's mark this fixed. |