I was testing the several locations for choosing node location. I've noticed the Tokyo location generating a lot of packets loss.

PING tokyo1.linode.com (106.187.33.12): 56 data bytes
Request timeout for icmp_seq 0
Request timeout for icmp_seq 1
64 bytes from 106.187.33.12: icmp_seq=2 ttl=54 time=262.138 ms
64 bytes from 106.187.33.12: icmp_seq=3 ttl=54 time=261.154 ms
64 bytes from 106.187.33.12: icmp_seq=4 ttl=54 time=282.449 ms
Request timeout for icmp_seq 5
64 bytes from 106.187.33.12: icmp_seq=6 ttl=54 time=321.101 ms
Request timeout for icmp_seq 7
64 bytes from 106.187.33.12: icmp_seq=8 ttl=54 time=367.544 ms
64 bytes from 106.187.33.12: icmp_seq=9 ttl=54 time=287.523 ms
Request timeout for icmp_seq 10
Request timeout for icmp_seq 11
Request timeout for icmp_seq 12
64 bytes from 106.187.33.12: icmp_seq=13 ttl=54 time=278.806 ms
Request timeout for icmp_seq 14
64 bytes from 106.187.33.12: icmp_seq=15 ttl=54 time=264.395 ms
Request timeout for icmp_seq 16
Request timeout for icmp_seq 17
64 bytes from 106.187.33.12: icmp_seq=18 ttl=54 time=261.331 ms
64 bytes from 106.187.33.12: icmp_seq=19 ttl=54 time=313.404 ms
Request timeout for icmp_seq 20
Request timeout for icmp_seq 21
64 bytes from 106.187.33.12: icmp_seq=22 ttl=54 time=262.614 ms
Request timeout for icmp_seq 23
64 bytes from 106.187.33.12: icmp_seq=24 ttl=54 time=260.117 ms
Request timeout for icmp_seq 25
64 bytes from 106.187.33.12: icmp_seq=26 ttl=54 time=269.424 ms
64 bytes from 106.187.33.12: icmp_seq=27 ttl=54 time=292.170 ms
Request timeout for icmp_seq 28
--- tokyo1.linode.com ping statistics ---
30 packets transmitted, 14 packets received, 53.3% packet loss
round-trip min/avg/max/stddev = 260.117/284.584/367.544/29.836 ms

It looks like the problems occur when the packets enter the ad.jp backbone.

traceroute to tokyo1.linode.com (106.187.33.12), 64 hops max, 52 byte packets
 1  cxxx.xxx.org (195.169.xxx.xx)  2.697 ms  1.793 ms  1.787 ms
 2  ae3.1503.jnr01.asd001a.surf.net (145.145.18.149)  1.946 ms  1.855 ms  1.894 ms
 3  ae1.500.jnr02.asd002a.surf.net (145.145.80.69)  2.272 ms  2.271 ms  2.047 ms
 4  xe-5-3-0.ams10.ip4.tinet.net (77.67.72.109)  2.549 ms  2.060 ms  1.956 ms
 5  xe-8-1-0.sjc10.ip4.tinet.net (89.149.183.17)  152.299 ms  153.155 ms  157.672 ms
 6  59.128.0.201 (59.128.0.201)  154.185 ms  153.990 ms  153.890 ms
 7  pajbb002.kddnet.ad.jp (124.211.34.17)  155.786 ms *  158.212 ms
 8  otejbb204.kddnet.ad.jp (203.181.100.201)  261.703 ms *  285.600 ms
 9  cm-fcu203.kddnet.ad.jp (124.215.194.180)  272.786 ms * *
10  * 124.215.199.122 (124.215.199.122)  268.269 ms *
11  * tokyo1.linode.com (106.187.33.12)  261.940 ms *

This should be fixed I guess...More users with same packets loss rates for Tokyo?

HOST: peregrine.asininetech.com   Loss%   Snt   Last   Avg  Best  Wrst StDev
  1. 212.111.33.229                0.0%    10    0.6   0.8   0.5   2.1   0.6
  2. 212.111.33.233                0.0%    10    0.5   0.6   0.5   0.7   0.1
  3. te3-1-border76-01.lon2.telec 40.0%    10    1.0   0.9   0.8   1.0   0.1
  4. 85.90.238.45                 30.0%    10    1.0   1.3   1.0   2.6   0.6
  5. xe-8-0-1.mpr2.lhr3.uskabove. 10.0%    10    1.6   1.0   0.8   1.7   0.4
  6. ge-4-2-0.mpr1.lhr3.uk.above.  0.0%    10    1.1   3.5   1.1  18.4   5.4
  7. xe-4-3-0.cr2.dca2.us.above.n  0.0%    10   75.5  75.5  75.4  75.8   0.1
  8. xe-0-2-0.cr2.iah1.us.above.n  0.0%    10  102.3 104.6 102.2 118.9   5.3
  9. xe-0-3-0.cr2.lax112.us.above  0.0%    10  163.3 138.3 133.9 163.3   9.9
 10. xe-1-1-0.mpr1.lax12.us.above  0.0%    10  133.8 141.7 133.8 184.2  17.3
 11. above-kddi.lax12.us.above.ne  0.0%    10  134.1 140.5 133.9 197.3  20.0
 12. lajbb002.kddnet.ad.jp         0.0%    10  138.6 143.4 138.5 171.6  11.0
 13. otejbb204.kddnet.ad.jp        0.0%    10  249.0 249.0 248.9 249.8   0.3
 14. cm-fcu203.kddnet.ad.jp       10.0%    10  255.3 257.3 253.1 264.5   4.4
 15. 124.215.199.122              10.0%    10  245.2 245.2 245.1 245.4   0.1
 16. tokyo1.linode.com            10.0%    10  245.0 245.4 244.9 246.6   0.7

That is a mtr from my London Linode.

no loss from my node in Fremont or from my office in southern California. Looks like you're both in Europe? Might be a routing issue from there?

 Host                                Loss%   Snt   Last   Avg  Best  Wrst StDev
 1. 184.105.143.85                    0.0%    12    0.5   0.7   0.4   3.1   0.8
 2. 10gigabitethernet2-3.core1.fmt1.  0.0%    12    9.7  10.1   0.4  57.1  17.4
 3. 10gigabitethernet1-1.core1.pao1.  0.0%    11    0.9   2.6   0.8  10.7   3.3
 4. 124.215.192.97                    0.0%    11    1.0   1.0   0.9   1.1   0.0
 5. pajbb001.kddnet.ad.jp             0.0%    11   17.3   6.8   0.8  37.9  11.8
 6. otejbb203.kddnet.ad.jp.100.181.2  0.0%    11  111.4 112.0 111.3 118.2   2.1
 7. cm-fcu203.kddnet.ad.jp            0.0%    11  121.9 116.9 112.6 123.9   4.2
 8. 124.215.199.122                   0.0%    11  110.5 110.3 110.3 110.5   0.1
 9. tokyo1.linode.com                 0.0%    11  113.0 112.6 112.5 113.0   0.1

|------------------------------------------------------------------------------------------|
|                                      WinMTR statistics                                   |
|                       Host              -   %  | Sent | Recv | Best | Avrg | Wrst | Last |
|------------------------------------------------|------|------|------|------|------|------|
|                415-vl109.ucinet.uci.edu -    0 |   13 |   13 |    0 |    0 |    0 |    0 |
|            cs1-core--415.ucinet.uci.edu -    0 |   13 |   13 |    0 |    0 |    0 |    0 |
|          kazad-dum-vl123.ucinet.uci.edu -    0 |   13 |   13 |    0 |    0 |    1 |    1 |
|         dc-tus-agg2--uci-ge-1.cenic.net -    0 |   13 |   13 |    0 |    0 |    1 |    1 |
|    dc-tus-agg1--tus-agg2-10ge.cenic.net -    0 |   13 |   13 |    0 |    0 |    1 |    1 |
|   dc-lax-core2--tus-agg1-10ge.cenic.net -    0 |   13 |   13 |    2 |    2 |    3 |    2 |
|  dc-lax-px1--lax-core2-10ge-2.cenic.net -    0 |   13 |   13 |    2 |    2 |    2 |    2 |
|                        peering.kddi.com -    0 |   13 |   13 |    2 |    3 |   22 |    2 |
|                   lajbb001.kddnet.ad.jp -    0 |   13 |   13 |    2 |    3 |   19 |    2 |
|                  otejbb203.kddnet.ad.jp -    0 |   13 |   13 |  102 |  102 |  103 |  102 |
|                  cm-fcu203.kddnet.ad.jp -    0 |   13 |   13 |  102 |  107 |  115 |  103 |
|                         124.215.199.122 -    0 |   13 |   13 |  110 |  110 |  110 |  110 |
|                       tokyo1.linode.com -    0 |   13 |   13 |  102 |  102 |  105 |  102 |
|________________________________________________|______|______|______|______|______|______|

mtr is better than traceroute.

An MTR (or trace) in one direction is kinda worthless. Need to see both paths - forward and reverse.

-Chris