New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Providers in Netherlands that don't use upstreams AS1299 neither AS174
nexius1981
Member
in Help
Good morning everybody!
Suddenly I found myself in a situation I never knew before...
I had 2 vps's with providers that the upstream goes through AS1299 and AS174.
Unfortunately, there are providers that no matter how much you show them the packet loss in these two tiers, they don't want to know and simply say it's not in their hands!
When they tell me that, there's really nothing they can do?!?!?!?!
Is there any provider that clearly does not use these AS1299 and AS174 upstreams in the Netherlands?
Comments
mostly that's impossible cause both arelion and cogent are tier1
How can the majority use services with such huge packet loss?
Where do you have the packet loss to ?
Can you show an MTR for a pair you have an issue with?
The issue might be with their Customer not upgrading, not with Arelion (which is a good network!) or Cogent.
https://imgur.com/a/IpgXX96
on this one only shows AS1299, but in random tests I found the same percenteage of packet loss on the upstream AS174
I start to feel this problem, when, inside VPN I had major packet loss between client to server or vice versa.
In some cases, even the handshake could not be made
In the first MTR, there is no packet loss to the destination (packet loss in between is irrelevant), and while in the second MTR is packet loss, it isn't routed over Arelion/Twelve99 or Cogent.
as174
https://imgur.com/a/YWD9qUm
SWEDEN REPRESENT
You can try our services via looking glass: https://www.hosteroid.uk/LGv2-new/
There’s no packet loss on this route either.
Look in practise.
If you are loosing 25% of packets on hop 2 then clearly hop 3 is expected to loose 25%+ and definitely destination has loss of 25%+.
You can’t loose 1/4 of the packets in the way but magically end up with 0% packet loss at the destination.
What you are observing is the control plane of the specific routers refusing to respond to your MTR packets which is normal and often expected. The packets traversing the router are not affected and consequentially you see 0% packet loss at the destination, so: there’s no packet loss.
irrelevant or not, the problem is that I had packet loss inside vpn and this packet loss lead to a lot of disconections due to handshakes not done. Tried Openvpn and wireguard and all this 2 protocols had problems.
This is the @Clouvider I like.
If you can't reproduce the package loss with any MTR, solving it won't be possible.
The cause was however most likely not the demonstrated paths you showed the MTRs for over Arelion.
I only have problems with those 2 VPS's that data pass trough this 2 upstreams...
the others VPS's that I have across UK and DE, non of them have this problems, and none of them show me any kind of packet loss on any upstreams
Ok, so get an MTR showing package loss. I don't know why you're not taking action based on the very good advice you got from @Clouvider and @lukast__ instead of going in circles about this.
Based on the MTRs you've shared, there simply isn't any packet loss to speak of. 8.8.8.8 (one of the test targets you used) ratelimits ICMP packets, so seeing packet loss on that destination is expected.
Are any of your services impacted by network issues, or did you just perform an MTR out of curiosity?
@CConner @Clouvider @emgh @lukast__
MTR from a vps that I have UK based to a vps in Netherlands
https://imgur.com/a/Sgr0ood
Still zero loss. Did you look at the screenshot yourself?
No loss; Did you see the explanation in my post above?
I aprecciate your answers @Clouvider and @emgh
Yes I read carefully tour answers. But what more tests can I do if I only have conectivity problems with 2 vps that are NL based with this 2 upstreams AS1299 AS174?!?!?
I'm in a middle of crossroads
You might be looking at this through the lens of survivorship bias. Are those 2 VPSs from the same provider? Any OpenVPN/Wireguard logs that you can look into and see what went wrong when the disconnection happens? All the above replies already explained the MTR results and it's unlikely that those upstreams are the cause of the issue you are encountering ..
Did you run an MTR between yourself and the VPS’es and the other way around? Since that’s where the issues are?
Exactly. If Cogent and Arelion had so much general package loss in NL that connecting to WireGuard isn’t reliably possible, there’s be tons of threads about it.
I have a feeling it’s a config issue or his local ISP.
https://lowendtalk.com/discussion/201406/ping-inside-my-vpn-route-are-getting-more-than-50-packet-loss#latest
yes, that is why I have opened a topic last week... cause I had massive packet loss inside my vpn
I didn't start doing this out of curiosity. I started to have a lot of disconnections within the VPN itself. When I did the basics of the basics, which is starting with ping, I saw a massive amount of lost packets. And I did this from several clients to the server.
Local ISP?!?!? @emgh I tested more than 10 VPN clients from various parts of the globe, from France, United Kingdom, Italy and Austria.
Demand justice for nexius1981!!!
Arelion and Cogent why do you ignore routing problems that don't exist! You evil bunch!
I look like an idiot among sharks. Don't get me wrong!
But in more than 10 VPS's that I have around the world, you may not believe it, but I only have problems in the 2 that I have in the Netherlands.
In the 2 vps in the Netherlands, I have different operating systems (Ubuntu and Debian)
When I started to dig, dig and dig I just checked a coincidence and with the help of the other topic, I verified that the coincidences were in these upstreams as174 and as 1299
All the others VPS's when I do MTR, all of them gave 0% packet loss on all routes!
Also, proper MTR should include both ways.. Server > Destination but also Destination > Server. So maybe problems you have with your VPN is actually on your ISP's side
from our SG IP to @nexius1981 's provider LG IP
Keep in mind that AS212477 / Royale Hosting has inline DDoS mitigation active in that point of presence. VPN traffic carried over UDP may be impacted by certain ratelimits / inspection rulesets, especially at higher packet rates. Support will most likely be able to tell you more and perhaps tweak some of the settings on your specific address(es).
DDoS inspection & mitigation is only performed on inbound traffic, which is why it's not showing up on outbound traceroutes, but is potentially impacting inbound traffic.