I would like to run a StorjNode via AirVPN using WireGuard. Currently, the MTU is set to 1320, and I have encountered issues with QUIC at that setting in the past. Does anyone have any experience with this? I know for a fact that an MTU of 1420 works for me, whereas 1320 has consistently resulted in faulty QUIC connections so far. Perhaps I am simply making a mistake somewhere.
Delete MTU line from the config. Delete DNS. Set explicit bind to the IP. QUIC works fine.
However with AirVPN performance and reliability is horrific. You will be better off running wireguard server on free Oracle instance.
Here is write up. Hosting services behind a restrictive firewall/CGNAT using DNAT on a VPS hosted wireguard endpoint | Trinkets, Odds, and Ends
Or better yet ask your ISP to forward a port to you — I assume you are behind CGNAT that’s why you are exploring VPN options?
I have a fiber-optic connection with a 300 Mbps upload speed, and I would like to run an additional Storj Node. I live in Germany, but I want to host this node in the USA—hence the need for a VPN connection. I plan to start with a test run in the USA; if that goes well, I intend to expand to Spain and other parts of Europe. Due to the /subnet rule, I will require WireGuard tunnels with dedicated exit IPs for those locations. This is the only provider I’ve found that offers few to no “neighbors” on the network.
If you are doing this to circumvent /24 rule — you are doing this to make correlated nodes look independent to the satellites. Don’t do it.
The /24 rule is a shorthand for node correlation and risk management. It exists for the reason — to allow operators run multiple nodes — so they can delete them separately when they want to reclaim space all while ensuring correlated nodes don’t receive disproportionate amount of traffic — because they functionally still behave like one node from the risk perspective.
Don’t do that.
Let me put it this way: If everyone does this, storj assumptions about node risk distribution will be violated, eventually it will lead to customer data loss, and either storj will effectively cease to exist, operators will be paid zero, or storj will start detecting such abuse, and wasting resources on mitigation, costing everyone. Neither is desirable outcome.
Don’t sell your integrity for $20/month.
I hadn’t looked at it that way, and it strikes me as logical. You are right about that.