All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
VPS VPN-ing trouble...?
Fizzy_LaFizz
Member
Hello all!
I have a VPS from OVH which I use as a VPN server to aid in connecting with a radio server. The radio server is located in an OVH facility in France (not sure which), while the VPS I have is in OVH Singapore.
I have set up the VPN successfully, following the steps in this tutorial https://www.digitalocean.com/community/tutorials/how-to-set-up-an-openvpn-server-on-ubuntu-18-04
Also note that the VPS is running Ubuntu 20.04
Now, here comes my problem. When doing a speedtest to the radio server while RDP'ed into the VPS, I'm getting 100+Mbps download. However, when I speedtest using the VPN on the VPS, I'm only getting about 3Mbps both ways.
Note that I'm not using Ookla's speedtest but rather a speedtest utility hosted on the radio server to measure my speeds specifically to the server.
This is what I'm getting when testing direct from the server (by virtue of RDP)
and this is what I'm getting when testing from client via VPN on the VPS:
The setup of the VPN using OpenVPN is done according to the tutorial I linked above. No other config changes were done accept the changes put forth in the tutorial.
How can I make it so that at least I have more decent speeds connecting to the radio server when connected via VPN? The entire reason why I set up the VPN in the first place is so that I can have better speeds to the server, but so far I'm getting negligible improvements, if any.
*I don't actually know if the screenshots are visible. This is my first time posting.
Comments
Pics not showing. The mobile image feature on Vanilla forum is garbage.
Also, you're only going to get better speeds if you're working around a bottleneck, otherwise you just added latency and buffering. Using the RDP will be better experience. 100Mbps to the VPN server means a max of 50Mbps to you since VPN server is doing 50Mbps each way (not exactly, but simple view is close enough).
Lastly, openvpn is old and slow, the cool kids use Wireguard. Changing to Wireguard alone should improve speeds. Enabling BBR should as well.
Also, I didn't check the openvpn instructions, but if you're using TCP instead of UDP, then that would be wrong.
Wireguard everything.
Welcome to LET and enjoy your stay
Yeah, there is somewhat of a bottleneck when connecting without a VPN. I am fully aware that I'm not and will never be getting better speeds, but 3Mbps compared to the actually decent speeds I could be getting is a joke.
I'll be looking into Wireguard and see if that helps, and if it does I'll update you guys.
Thank you!
Sounds like you should blame your local connection to VPN server and/or VPN configuration.
Try WG anyway, mostly it will work. If still slow, consider OpenVPN TCP mode and get a traceroute result to see if your traffic is being routed through elsewhere.
UDP Traffic is throttled at OVH, however, it should not be that low.
Try Wireguard if the cpu usage is to high.
Likely your ISP seems to be the issue.
Okay so update:
(I am aware this is probably a very unsecure way of setting it up. Bear with me. I'm still learning the ropes.)
I have set up Wireguard, and can connect to it...but when I do, I can only access google.com. Nothing else works. I need some help here. Followed tutorial here:
If I google "what's my IP" when connected to the VPN, it does show the IP of my VPS and not my local machine.
I have no idea what's wrong, I have no idea how to fix it, and I have been prodding around since yesterday afternoon. I am frustrated.
Could anyone of you perhaps help me out?
I don't get how my ISP could be the issue. When I connect using OpenVPN from my phone, which is connected to the same internet connection as my laptop, I'm getting better speeds over WiFi.
You need to do this scientifically instead of throwing yourself randomly at all the things people suggest, you immediately blamed and ditched OpenVPN for WG at a moment's notice.
I do know WG has better performance across the boards, but OpenVPN shouldn't have problems handling a few Mbps of traffic.
1) I believe you still haven't ruled out a poor link between you and your VPS. Try running
iperfon your VPS then check if you get around 100Mbps (or any sane value) without the VPN.2) Now repeat the previous step, but now measure from within the VPN/WG tunnel.
If that was the problem, then you can repeat the test again using another tunnel; try with OpenVPN, WG or anything that tickles your fancy.
If you got good speeds from both previous items, then you can probably rule out a problem between you and the VPS.
3) Check to make sure your requests to the radio station are actually being routed through the VPN (you may have forgotten a static route you set up before or something else is interfering with the routing). While connected to the VPN, try
traceroute-ing to the radio station and check if your traffic does go inside the tunnel.I have done all three things you told me to. iperf-ing gives me a very good result to the VPS, basically maxing out my broadband connection from the ISP, both connected AND disconnected to/from the VPN.
Tracerouting to the radio server while connected to the VPN shows that my traffic is being routed via the tunnel. Indeed, when I traceroute from the VPS itself to the radio server, it's showing more or less the same route as the traceroute from my PC.
This makes me wonder even more, what could be the barrier? I'm still getting only 3Mbps to the radio server when connected to the VPN.
This was all done via OpenVPN btw. I still can't get WG to work (well, I did, but I could only go on google.com and nothing else)
Aside from speed test being slower, what else is affected?
Is your radio usage affected in any noticeable way?
How about doing an iperf from your machine to an arbitrary server through the VPN? Preferably if you can find one that's hosted somewhere near the radio server is at.
Nearing tinfoil hat territory but could be fucky routing/NAT-ing from within VPN VPS.
Well, I can't say I'm noticing any difference except for the radio server in that I'm having less ping but the same or slower speeds connecting to the radio server when connected via VPN.
Oh and naturally region bypass—I'm now getting Singaporean ads instead of the usual ads I see 😂
Well, I did an iperf to a random server in Denmark and I'm getting 17Mbps with 10 parallel streams. Still, much better than the 3Mbps I'm getting to the radio server.
Well I'm out of ideas, is this radio server public? Would like to offer to test to see if I get the same behavior
I don't think I'm allowed to disclose the test utility to the public. Thanks for the offer though.
I'm going to call this one a bust. I've done some arbitrary testing myself and "concluded that the test utility is fucked somehow and that I'm actually getting more speeds than indicated. This is backed up by my speedtesting (Ookla) to a nearby server with and without the VPN, and the results indicated that the speeds with the VPN connection is much greater (10 Mbps compared to 2 Mbps)."*
*this is probably just a lie, and my way of saying, "I've given up but I have my doubts anyway." But hey, at least I have some sort of basis of saying that with the VPN, I'm getting more speeds to continental Europe in general than without!
Thanks to everyone who helped, really appreciate it!
*also if there's something I should do after I deem a thread is finished, like somehow closing it, please tell me lol
Oh well, that's one more marble for the "unsolved mysteries" bag.
You can flag the post and request a mod to close the thread, but keeping it open isn't a problem either.