Open eloquentbit opened 4 years ago
I'm not sure if it's related, I've had DNS resolution issues since a couple of days as well.
Whenever dnscrypt-proxy choses the cloudflare
server, name resolution doesn't work. When I force it to use the cloudflare-ipv6
server, everything works as expected.
To force IPv6, set ipv4_servers = false
in /etc/dnscrypt-proxy/dnscrypt-proxy.toml
and restart dnscrypt-proxy via systemctl restart dnscrypt-proxy.service
.
Another solution for me was to follow these steps from the FAQ to use the Google DNS server. Might indicate an issue with the Cloudflare IPv4 DNS connectivity?
In my case, it's reproducible. Would be very interested if it helps you out as well.
Hi @oh-fv, thanks for the suggestions.
On my Ubuntu server there isn't /etc/dnscrypt-proxy/dnscrypt-proxy.toml
file, maybe because I have opted out for DNS encryption during installation.
I've tried also to follow the FAQ you have mentioned but without success.
I've tried both methods, tried rebuilding the server and even tried switching from cloudflare to nextdns in /etc/dnscrypt-proxy/dnscrypt-proxy.toml
nothing works.
My solution only worked for a couple of days for me. I did a fresh install yesterday evening (with DNS encryption and ad blocking), and after around 9 in the morning today DNS resolution stopped working again. Frustrating.
Is there anything I can do to help debug this issue? I'm running Algo on a fresh Ubuntu 18.04 install.
I did a 2nd re-install without any DNS features and things are working for me now. I would love to help debug this as well.
In an otherwise unchanged /etc/dnscrypt-proxy/dnscrypt-proxy.toml
I edited server_names
to server_names = ['cloudflare']
with the following reproducible result in # journalctl -u dnscrypt-proxy
:
Mar 20 21:19:18 v220190910411597575 dnscrypt-proxy[4202]: Get https://dns.cloudflare.com/dns-query?ct=&dns=yv4BAAABAAAAAAABAAACAAEAACkQAAAAgAAAAA: net/http: request canceled while waiting for connection (Client.Timeout exceeded while awaiting headers)
Mar 20 21:19:18 v220190910411597575 dnscrypt-proxy[4202]: dnscrypt-proxy is waiting for at least one server to be reachable
So for me, the issue still seems to be related to Cloudflare. I don't know enough about the inner workings of the Algo server to say if this is the cause or a symptom. But I picked a couple of servers from https://dnscrypt.info/public-servers/ I like and added them to server_names
(without the Cloudflare ones) and so far the DNS resolution works.
Chiming in here to say that I'm running into this issue now, too. Not sure what's up; it was working for a few days just fine last week, and now I can't create a fresh instance and then actually resolve from a Catalina 10.15.4 client. This is from Lightsail, configured with an Ubuntu 18.04.4 instance. Doesn't matter whether dns_encryption
is true or false in config.cfg
before running $ ./algo
.
Now I can't replicate. DNS resolves just fine. Not sure if it was a temporary hiccough or some configuration option I missed somewhere.
I can reproduce on commit 4172dea43659818fbbcba3ecc5e6bb54a0b98d72. Everything y'all have mentioned seems same for me. The only difference is that I am running this on Debian 10. I only modified the ubuntu.yml tasks to not update cache or use apt to install anything, and I installed wireguard, dnscrypt-proxy and co myself.
Fixed by setting the DNS
property in local client wireguard configs to 8.8.8.8. Seems to work for 1.1.1.1 as well.
Fixed by setting the
DNS
property in local client wireguard configs to 8.8.8.8. Seems to work for 1.1.1.1 as well.
Thanks for this. I don't really know what happened, but manually setting wg0.conf
on my client to 1.1.1.1 indeed worked. Thank you!
Maybe this will help someone, I had the same issue, turns out docker created a bridge interface for one of the containers which overlapped with the dns ip. Removing the interface (for example docker network prune
) solved it.
Describe the bug
I've deployed Algo in a brand new Vultr instance from my Mac OS Catalina. The connection is established without any problem but I can't resolve any hostname. From what I see, when connected, my DNS server in /etc/resolv.conf changes from my router's IP address to the address of wg0 interface (10.19.49.1) even if in config.cfg are present these settings: dns_encryption: false dns_servers: ipv4:
To Reproduce
Steps to reproduce the behavior:
Expected behavior
I'm expected to be able to perform DNS queries without errors
Additional context
Full log