Packet loss e0-30.core2.lon3.he.net 80%+

[invux]

Hi,

Is anyone else experiencing connection problems through the London PoP ?

I am getting 85% packet loss, which causes connection problems:

Code Select Expand

2. tunnel55578.tunnel.tserv5.lon1.ipv6.he.net                                    0.6%  1256    9.5  13.7   8.6  91.3  12.5
3. e0-19.core2.lon2.he.net                                                       2.1%  1256    8.8  12.2   7.8  95.2  11.1
4. (waiting for reply)
5. e0-30.core2.lon3.he.net                                                      86.1%  1256    8.6  16.7   8.1  88.6  17.0


This has been going on a while, 10 days or so.
Thanks.

[bmhughes]

I'm not seeing packet loss but am seeing high latency to tserv1.lon2.he.net

Code Select Expand


13. port-channel8.core2.lon2.he.net                                                                                                       5.2%    97   21.8  21.5  15.6 106.8   9.6
14. tserv1.lon2.he.net                                                                                                                    0.0%    97  152.4 157.9  44.5 345.0  50.9


Seem to jump once inside HEs network.

[tc424]

Seeing significant problems now reaching tserv1.lon2.he.net from a Sky/NOW! connection on v4:

Code Select Expand

4. 10gigabitethernet5-1.core1.lon1.he.net                                                            69.6%    24    6.0   6.1   5.9   7.3   0.5
5. 40ge1-3.core1.lon2.he.net                                                                        82.6%    24    6.9  23.0   6.9  31.8  11.6
6. port-channel8.core2.lon2.he.net                                                                   0.0%    24    7.8   7.6   7.3   8.9   0.3
7. tserv1.lon2.he.net                                                                                0.0%    24  245.0 273.7 186.3 410.9  55.1

Can anyone at he.net comment - is there anyway to raise this with anyone?

S.

[bmhughes]

Mine's been unusable for a couple of weeks now from virgin media. Has been solid for several years before this.

[bmhughes]

I've tested with another tunnel to the server at 216.66.80.26 (listed at tserv1 but shows at tserv5?) and the problems go away so it looks like an issue with that tunnel server. Annoying to change though as you can't move prefixes between servers.

[snewbury]

tserv1.lon1.he.net has started dropping about 5-15% ipv6 packets for the last few weeks.  I thought it might be Virgin Media too, but I set up a VM at Linode and pinged the other side of the tunnel with ipv6, which also resulted in packet loss.

PING 2001:470:0:67::2(2001:470:0:67::2) 56 data bytes
64 bytes from 2001:470:0:67::2: icmp_seq=1 ttl=58 time=0.781 ms
64 bytes from 2001:470:0:67::2: icmp_seq=2 ttl=58 time=35.4 ms
64 bytes from 2001:470:0:67::2: icmp_seq=3 ttl=58 time=0.879 ms
64 bytes from 2001:470:0:67::2: icmp_seq=4 ttl=58 time=0.835 ms
64 bytes from 2001:470:0:67::2: icmp_seq=5 ttl=58 time=84.7 ms
64 bytes from 2001:470:0:67::2: icmp_seq=6 ttl=58 time=0.833 ms
64 bytes from 2001:470:0:67::2: icmp_seq=7 ttl=58 time=0.857 ms
64 bytes from 2001:470:0:67::2: icmp_seq=8 ttl=58 time=0.859 ms
64 bytes from 2001:470:0:67::2: icmp_seq=9 ttl=58 time=0.935 ms
64 bytes from 2001:470:0:67::2: icmp_seq=10 ttl=58 time=0.890 ms
64 bytes from 2001:470:0:67::2: icmp_seq=11 ttl=58 time=0.846 ms
64 bytes from 2001:470:0:67::2: icmp_seq=12 ttl=58 time=0.886 ms
64 bytes from 2001:470:0:67::2: icmp_seq=13 ttl=58 time=0.889 ms
64 bytes from 2001:470:0:67::2: icmp_seq=14 ttl=58 time=0.878 ms
64 bytes from 2001:470:0:67::2: icmp_seq=15 ttl=58 time=0.856 ms
64 bytes from 2001:470:0:67::2: icmp_seq=17 ttl=58 time=0.882 ms
64 bytes from 2001:470:0:67::2: icmp_seq=18 ttl=58 time=0.844 ms
64 bytes from 2001:470:0:67::2: icmp_seq=19 ttl=58 time=0.893 ms
64 bytes from 2001:470:0:67::2: icmp_seq=20 ttl=58 time=0.935 ms
^C
--- 2001:470:0:67::2 ping statistics ---
20 packets transmitted, 19 received, 5% packet loss, time 19054ms
rtt min/avg/max/mdev = 0.781/7.094/84.660/19.834 ms

Also some big spikes there.  Some sort of load issue on tserv1.lon1.he.net?

[tho]

I'm seeing that too on ams1 tunnel.

Code Select Expand

5. par-th2-sbb1-nc5.fr.eu            0.0%   120    4.3   4.4   4.2   4.7   0.1
6. 10.200.2.71                       0.0%   120    4.5  97.5   4.2 606.6 133.4
7. he.par.franceix.net              97.5%   120    4.5  32.5   4.5  87.8  47.9
8. amsix-200gbps.core1.ams1.he.net  82.4%   120   76.8  40.0   9.6 175.0  48.4
9. tserv1.ams1.he.net               24.2%   120    9.8  11.2   9.4  88.2   8.6

The tunnel is pretty much unusable. It started maybe one month ago and I never saw that for 5 years. This is strange that several tunnels are affected since about the same time?

I opened a ticket one week ago but not getting much replies or insight.

[drsox]

I've seen months and months of ~6% packet loss between RapidSwitch or Andrews & Arnold and tserv5.lon1.ipv6.he.net and tserv1 which is making things unusable too.

What is frustrating is my server is in a part of their network where they supposedly can't support native IPv6; so I'm having to move servers because the tunnel is now no longer reliable enough.

Graphs over a few days time and traceroutes attached.

Given how long it has been going on.. I feel like it's unlikely to get fixed.

Note: I do not see the same packet loss to tserv1.lon2.he.net; so it's just lon1 which I see a problem to.

[lewislfoster]

For the past few months ive been seeing consistent 5-10% packet loss to e0-19.core2.lon2.he.net both through tunnel broker and native IPv6 from linode. Before i tried to ping from linode i thought it might have been something to do with virgin media but it appears to not be