Fix openvpn server side deployment

[hellais]

In #195 the greenhost deployment of the openvpn server was terminated.

It was not clear to us at the time that this would impact probes, but it did since that was the only true server running openvpn, while the other endpoints were configured to just be simple TCP proxies.

As a result of this we are now in a situation where all openvpn tests are failing as the proxying service is not forwarding traffic to any valid backend.

We should find a way to unbrick this by:

• Setting up a new instance of the openvpn service
• Migrating the keys over to the new service (this is pending on getting the private keys from @ainghazal as we don't have a backup of them)
• If we aren't able to migrate they keys we will have to create new credentials and ship an update of probe with the new keys

[hellais]

I have setup a new instance of this service on hetzner. It has new keys inside of it which we are going to add to the new release of the engine.

We would still like to get a copy of the older keys so that we are able to restore functionality of older probes that were using that.

[hellais]

We weren't able to get the deployment of the new instance to work with OONI Probe, as we see errors in the client and the server:

client side error:

[     33.502669] <info> [@] S_PRE_START -> S_PRE_START
[     33.502790] <debug> > P_CONTROL_HARD_RESET_CLIENT_V2 {id=0, acks=[]} localID=d89848e31c08492e remoteID=0000000000000000 [0 bytes]

server side error:

2025-03-31 21:04:31 client-cert-001/109.40.243.185:3628 Data Channel: cipher 'AES-256-GCM', peer-id: 0, compression: 'stub'
2025-03-31 21:04:31 client-cert-001/109.40.243.185:3628 Timers: ping-restart 10
2025-03-31 21:04:32 read UDPv4 [ECONNREFUSED]: Connection refused (fd=5,code=111)
2025-03-31 21:04:36 read UDPv4 [ECONNREFUSED]: Connection refused (fd=5,code=111)
2025-03-31 21:04:41 client-cert-001/109.40.243.185:3628 [client-cert-001] Inactivity timeout (--ping-restart), restarting
2025-03-31 21:04:41 client-cert-001/109.40.243.185:3628 SIGUSR1[soft,ping-restart] received, client-instance restarting

@ainghazal mentioned this might have to do with the wrong version of the openvpn server being used inside of the image we build as part of: https://github.com/ooni/docker-openvpn-test.

[hellais]

Given that we don't have a way of reliably redeploying the openvpn serverside component and everything currently deployed doesn't work, we have disabled this test by setting the domain name records for them to 127.0.0.42 which has the effect of disabling the test.