[Myatu]

Try sending a message to oles@ovh.net...

[Emx]

The old route trough interoute returned. Are you doing ANYTHING to return the old, best route trough decix.telekom.rs? Millions of users are affected by low speeds and damn high latency (ping)!

I've sent mails to noc@ovh.net, customersupport@ovh.co.uk and after 7 days the situation is still the same.

[Emx]

You can lock this topic. Great technical team at OVH fixed the route and it's 40 ms now!

1 * * *
2 vss-1-6k.routers.chtix.eu (91.121.131.106) 120.440 ms * *
3 20g.fra-1-6k.routers.chtix.eu (91.121.131.85) 8.443 ms * *
4 decix.telekom.rs (80.81.192.195) 33.228 ms 35.521 ms 35.518 ms
5 212.200.227.246 (212.200.227.246) 33.171 ms 31.003 ms 31.007 ms
6 79-101-96-98.static.isp.telekom.rs (79.101.96.98) 36.761 ms 40.015 ms 36.192 ms
7 81.93.65.82 (81.93.65.82) 38.816 ms 36.082 ms 36.682 ms
8 81.93.65.174 (81.93.65.174) 39.409 ms 42.054 ms 41.524 ms
9 81.93.66.254 (81.93.66.254) 42.915 ms 40.145 ms 42.134 ms
( telekom of republik of srpska, now it goes to smaller isps after this.)

Great!!

[Psihonaut]

I am from Bosnia and I can just confirm that many are complaining about high pings to servers in western Europe. At first I thought it was fault of my ISP, but then I heard that situation is same even in neighbour country, Serbia. I would say it started at beginning of February and its still actual. I hope someone(don't know who is actually responsible for that routing ) will eventually take care about that because millions of users are experiencing problems.

[Emx]

Any updates on this, any ticket @ travaux.ovh.net? I don't see any improvemenst in routing/connection/latency

[Emx]

Ah thanks, thank you for great support, keep me informed in this topic when you get updates from network team

[Neil]

Despite the problem is in the routing not in the first or second hop / particulary ES/FR? So just post this topic to .es forum right

Your server is with OVH.es, we can help but we do not support these servers. But I have already asked the network team, but in future you need to contact them.

[Emx]

Hi

I have heard back and the routing will be fixed when it is off-peak, however your server is with OVH.es so this issue should be put to the attention of them.

Despite the problem is in the routing not in the first or second hop / particulary ES/FR? So just post this topic to .es forum right

[Neil]

Neil, my server is in France/Roubaix but bought trough .es. It doesn't matter really since the problem in the routing is not in France. This problem didn't actually appear week or two ago but it exists for longer time. And the current route to SE europe, particulary ex-yugoslavia I'm talking about, is not even approximately optimal.

keep me updated, thanks.

Hi

I have heard back and the routing will be fixed when it is off-peak, however your server is with OVH.es so this issue should be put to the attention of them.

[Emx]

Btw, amis ISP in Croatia (www.amis.hr) is very decent example of what I'm talking about. Look at these nice latency and short hops:

1 188.165.194.252 (188.165.194.252) 43.917 ms * *
2 20g.fra-1-6k.routers.chtix.eu (91.121.131.85) 17.956 ms * *
3 30g.pra-1-6k.routers.chtix.eu (213.251.128.109) 25.911 ms * *
4 20g.vie-1-6k.routers.chtix.eu (213.251.128.98) 30.554 ms * *
5 vix.amis.net (193.203.0.117) 31.012 ms 31.249 ms 31.242 ms
6 zagreb6-ge-9-16.amis.net (212.18.41.166) 37.429 ms 29.544 ms 31.562 ms
7 hosting.amis.hr (194.146.109.202) 36.766 ms 36.991 ms 36.993 ms

cool, nah? roubaix->frankfurt->prague->vienna->zagreb.

[Emx]

Neil, my server is in France/Roubaix but bought trough .es. It doesn't matter really since the problem in the routing is not in France. This problem didn't actually appear week or two ago but it exists for longer time. And the current route to SE europe, particulary ex-yugoslavia I'm talking about, is not even approximately optimal.

keep me updated, thanks.

[pdu]

I don't know if it's directly relevant here as I haven't the time to investigate your routes specifically, but OVH have been doing some maintenance on a number of routers over the last few days, including a few issues affecting a couple of their major cisco routers where they found a bug after a change, so whilst there might be a more optimal route even after it's all completed and settled down again, I suspect some of your recent issues are related to that. I too have had some issues with latency on a couple of my servers over the weekend (even low bandwidth stuff like ssh hanging for a few seconds) but others have been as speedy as ever. Fingers crossed all the hard work from the network guys pays off and solves all the latency speed issue posts we've seen in here over the last few days

P.s. to staff, used virtual mac functions again today on a new server, worked perfectly first time, thank you so much

[Neil]

Hi

Thank you for the feedback, I have passed it on. But you can also get the owner of the server to contact the spanish support about this issue.

[Emx]

Hello! Can you (OVH' team, Oles personally? ) please please investigate the routing from OVH's servers (Roubaix, or generally) to south-east Europe? One major ISP in here uses DTAG to approach OVH's servers. Other traceroute to other ISP gave me these details, showing some problems in interoute's connections and many hops:

1 188.165.194.252 (188.165.194.252) 0.658 ms * *
2 20g.ldn-1-6k.routers.chtix.eu (94.23.122.65) 4.142 ms * *
3 * * *
4 * * *
5 * * *
6 * ae0-0.ams-koo-score-2-re1.interoute.net (84.233.190.2) 27.916 ms 28.188 ms
7 ae1-0.fra-006-score-1-re1.interoute.net (84.233.190.50) 28.210 ms 28.211 ms 28.173 ms
8 ae1-0.vie-per-score-1-re0.interoute.net (212.23.43.25) 28.172 ms 28.171 ms 28.171 ms
9 ae0-0.vie-per-score-2-re0.interoute.net (212.23.43.50) 38.271 ms 38.278 ms 38.277 ms 10 ae1-0.bts-001-score-1-re1.interoute.net (84.233.147.13) 28.156 ms 28.156 ms 28.156 ms
11 ae0-0.bts-001-score-2-re1.interoute.net (84.233.147.2) 28.162 ms 28.162 ms 27.929 ms
12 ae1-0.bud-001-score-2-re1.interoute.net (84.233.147.113) 28.099 ms 28.094 ms 28.089 ms
13 Gi2-0.bud-001-access-300.interoute.net (84.233.147.122) 28.084 ms 28.080 ms *
14 84.233.170.166 (84.233.170.166) 203.982 ms 203.986 ms 203.775 ms

ERROR appears in the upper - notice the 200ms higher lag? Final ping increases up to 250+ ms and should be 60ms at maximum, as I've experienced it.
I've tested this several times for few days and it gets laggy afternoon.

Route goes like this: OVH (Roubaix) -> London -> Amsterdam -> Frankfurt -> Vienna -> Budapest.

Better route which I propose and would be better to use, which would give more speed and less hops and latency, would be to use:
OVH -> London -> Amsterdam -> DTAG (Deutsche Telekom) in Germany -> Croatia / Serbia where are my ISP's sub providers.

Or even use OVH -> Frankfurt -> DTAG -> Croatia / Serbia (T-COM or Telekom Srbije).
Like this:
1 44 ms 45 ms 50 ms nsXXXXXX.ovh.net [188.165.194.XXX]
2 62 ms * 228 ms 20g.vss-2-6k.routers.chtix.eu [94.23.122.117]
3 * 47 ms * 20g.fra-1-6k.routers.chtix.eu [91.121.131.70]
4 44 ms 44 ms * 20g.dtag.fra-5-6k.routers.cht [213.251.130.25]
5 40 ms 38 ms 38 ms 217.239.40.222
6 26 ms 25 ms 24 ms 194.25.209.85
7 13 ms 13 ms 25 ms gvz01-gvz11.ip.t-com.hr [195.29.241.46]
8 13 ms 14 ms 14 ms gvz11-hdr01.ip.t-com.hr [195.29.241.90]
9 15 ms 14 ms 15 ms hdr01-gdr10.ip.t-com.hr [195.29.240.94]

etc. In case of T-COM (Croatia). See? Good ping at least from my side.

Also there is this route directly trough GBLX:

1 188.165.194.252 (188.165.194.252) 7.531 ms * *
2 * * *
3 80g.th2-1-6k.routers.chtix.eu (213.186.32.166) 132.791 ms * *
4 80g.gsw-2-6k.routers.chtix.eu (213.186.32.162) 7.368 ms * *
5 10g.gblx.gsw-2-6k.routers.chtix.eu (213.186.32.198) 4.502 ms 4.737 ms 4.730 ms
6 CROATIAN-TELECOM.POS8-0-0.ar2.VIE1.gblx.net (64.215.185.194) 49.476 ms 49.264 ms 49.671 ms
7 GUT04-GTR02.ip.t-com.hr (195.29.249.90) 49.645 ms 49.102 ms 52.183 ms^

And to Telekom Srbija:

1 188.165.194.252 (188.165.194.252) 0.427 ms * *
2 * * *
3 80g.th2-1-6k.routers.chtix.eu (213.186.32.166) 62.418 ms * *
4 80g.gsw-2-6k.routers.chtix.eu (213.186.32.162) 4.419 ms * *
5 10g.gblx.gsw-2-6k.routers.chtix.eu (213.186.32.198) 31.744 ms 31.737 ms 31.728 ms
6 TELEKOM-SRBIJA.s4-0-0.ar1.VIE1.gblx.net (64.210.69.2) 125.124 ms 152.645 ms 120.959 ms
7 212.200.6.97 (212.200.6.97) 121.214
Direct routing OVH->GBLX uplink of Telekom Srbija in Vienna->Belgrade. Sounds good Altought it's currently large ping. :/

Telekom Srbija has links to DECIX and GBLX and LEVEL3. You can check and see if there's better connection/speed and lower latency using OVH's peering with DECIX/GBLX/LEVEL3 if there's any, or those links regardless if it's peering or transit.
T-COM has links to Telianet, interoute and DTAG.

Interoute is the worst solution in both cases - large route and high pings.


Please check this since it's nonsense to use so many hops and just going into circle when there are direct OVH's connection from France to Frankfurt and then from Germany to Croatia / Serbia. C/S routers will handle the rest routing to my ISP.

This issue is relevant to more than 40 ISPs and more than 3 millions Internet users.
Use your own infrastructure and do not route to London and Amsterdam which is not necessary at all.

Since I'm OVH's customer more than one year and use not the basic server but the more expensive due to size of my business and it's purpose, I've always been happy with the performance and the service but I think there can some fine-tunings/adjustments.

Please check and report as soon as possible, I would like to talk to tech-staff.

Thanks for great service so much time despite some recent increases in price - I'm fine with it. Just lower the latency/hops.

And appreciate my not so short, very descriptive post. I can provide more technical information if necessary.