Forums » Bugs
Packet loss on the intertubes near VO's servers.
This might explain why a lot of people have been reporting problems connecting and getting kicked off lately.
I've run an MTR from home and from my shellbox and both are seeing ~25% packet loss at ae-22-52.car2.Chicago2.Level3.net - the host right before redanvil's router.
I've run an MTR from home and from my shellbox and both are seeing ~25% packet loss at ae-22-52.car2.Chicago2.Level3.net - the host right before redanvil's router.
Yeah pretty much
Traceroute to majikthise:
[..]
6 xe-11-0-0.bar2.boston1.level3.net (4.79.2.89) 24.254 ms 25.612 ms 22.796 ms
7 ae-0-11.bar1.boston1.level3.net (4.69.140.89) 27.665 ms 22.942 ms 24.607 ms
8 ae-5-5.ebr1.chicago1.level3.net (4.69.140.94) 50.844 ms 80.114 ms 43.470 ms
9 ae-6-6.ebr1.chicago2.level3.net (4.69.140.190) 73.660 ms 46.040 ms 58.215 ms
10 ae-12-51.car2.chicago2.level3.net (4.69.138.133) 44.807 ms * *
11 red-anvil-l.car2.chicago2.level3.net (4.30.14.162) 54.289 ms 55.896 ms 57.566 ms
[..]
Traceroute to majikthise:
[..]
6 xe-11-0-0.bar2.boston1.level3.net (4.79.2.89) 24.254 ms 25.612 ms 22.796 ms
7 ae-0-11.bar1.boston1.level3.net (4.69.140.89) 27.665 ms 22.942 ms 24.607 ms
8 ae-5-5.ebr1.chicago1.level3.net (4.69.140.94) 50.844 ms 80.114 ms 43.470 ms
9 ae-6-6.ebr1.chicago2.level3.net (4.69.140.190) 73.660 ms 46.040 ms 58.215 ms
10 ae-12-51.car2.chicago2.level3.net (4.69.138.133) 44.807 ms * *
11 red-anvil-l.car2.chicago2.level3.net (4.30.14.162) 54.289 ms 55.896 ms 57.566 ms
[..]
ARF: are you seeing the loss propagating past that router? Or just at that router? It's pretty common for edge routers to de-prioritize directed ICMP responses, as mentioned here.
Ie, unless you're seeing 25% packet loss to us, this measurement of an issue at the L3 router is spurious and effectively useless.
I'm all for debugging any problems, and if we can get some conclusive data I'll push it upstream to RA, who will then talk to Level3. But we'll need something more concrete.
Ie, unless you're seeing 25% packet loss to us, this measurement of an issue at the L3 router is spurious and effectively useless.
I'm all for debugging any problems, and if we can get some conclusive data I'll push it upstream to RA, who will then talk to Level3. But we'll need something more concrete.
Just that hop unfortunately, which I know isn't really that useful, but there's also the in-game effects like what slime posted.
Right, and I'd really like to reduce the instances of the stuff that slime posted, but the reported level3 spike is almost certain to have absolutely nothing to do with the game issues.
So, when the actual packet loss or network issues are occurring, that is the best time to try and do some sort of measurement. If you can get me some information on where the problem might be, I'll try to escalate it (with very limited enthusiasm for success).
So, when the actual packet loss or network issues are occurring, that is the best time to try and do some sort of measurement. If you can get me some information on where the problem might be, I'll try to escalate it (with very limited enthusiasm for success).
Sporadic network problems are an absolute JOY to try and nail down. Especially when a bunch of other people's routersare involved that may or may not be filtering/de-prioritizing/QoS-ing/etc.
Is there any way in the game client to do a network packet loss graph or totals like the unix 'ping' command does? I somehow doubt the server will respond if I just start sending UDP at it from hping3. :)
Is there any way in the game client to do a network packet loss graph or totals like the unix 'ping' command does? I somehow doubt the server will respond if I just start sending UDP at it from hping3. :)