spantaleev
Trefik unable to resolve matrix-container-socket-proxy dns
Describe the bug
After upgrading traefik cannot resolve the ip address for the container matrix-container-socket-proxy as seen by the following error logs.
Dec 25 07:16:40 matrix.example.lan matrix-traefik[2458633]: time="2023-12-25T12:16:40Z" level=error msg="Provider connection error error during connect: Get \"http://matrix-container-socket-proxy:2375/v1.24/version\": dial tcp: lookup matrix-container-socket-proxy on 127.0.0.11:53: server misbehaving, retrying in 26.521230813s" providerName=docker
Dec 25 07:16:43 matrix.example.lan matrix-traefik[2458633]: 10.0.1.84 - - [25/Dec/2023:12:16:43 +0000] "GET /_matrix/client/v3/sync?full_state=false&timeout=30000&since=s40044_1494058_26_17202_36448_16_9656_168_0_1 HTTP/1.1" 404 19 "-" "-" 81 "-" "-" 0ms
Dec 25 07:16:46 matrix.example.lan matrix-traefik[2458633]: 71.27.185.41 - - [25/Dec/2023:12:16:46 +0000] "GET /_matrix/client/versions HTTP/2.0" 404 19 "-" "-" 82 "-" "-" 0ms
Dec 25 07:16:47 matrix.example.lan matrix-traefik[2458633]: 10.0.1.3 - - [25/Dec/2023:12:16:47 +0000] "HEAD / HTTP/2.0" 404 19 "-" "-" 83 "-" "-" 0ms
Dec 25 07:16:49 matrix.example.lan matrix-traefik[2458633]: 10.0.1.84 - - [25/Dec/2023:12:16:49 +0000] "GET /_matrix/client/v3/sync?full_state=false&timeout=30000&since=s40044_1494058_26_17202_36448_16_9656_168_0_1 HTTP/1.1" 404 19 "-" "-" 84 "-" "-" 0ms
Dec 25 07:16:50 matrix.example.lan systemd[1]: run-docker-runtime\x2drunc-moby-1f88275785b3c6aa7bf2396ea8db8a166a651fbfe0afc4007f978ef368f753e1-runc.zb8jxa.mount: Deactivated successfully.
Dec 25 07:16:50 matrix.example.lan systemd[1]: run-docker-runtime\x2drunc-moby-35b2c3ecdb06e5daf7b2ca25b34406bc6cb37eda243597a27e90750750f9958e-runc.rpVMOn.mount: Deactivated successfully.
Dec 25 07:16:53 matrix.example.lan matrix-traefik[2458633]: 71.27.185.41 - - [25/Dec/2023:12:16:53 +0000] "GET /_matrix/client/versions HTTP/2.0" 404 19 "-" "-" 85 "-" "-" 0ms
After troubleshooting this appears to be because traefik and were not on the same network. matrix-container-socket-proxy was on its own network by itself and traefik was on the traefik network.
This occured for me when upgrading from commit 3dc77c4f9a1277f2338f2cd3664186fd94b93292 ( Synapse 1.96.1 ) to commit da27655ef34999fa924bc0a5e641dbd9ba06f133 [December 23, 2023 Merge pull request] (https://github.com/spantaleev/matrix-docker-ansible-deploy/pull/3060 from etkecc/fix-chatgpt-auth)
I filed a ticket in the role as well at Trefik unable to resolve matrix-container-socket-proxy dns . I am writing it in this ticket in case others face the same issue. I believe this is an issue in the traefik role.
Work Around Set the following in my matrix variables:
devture_container_socket_proxy_enabled: false
Matrix Server:
- OS: Ubuntu 22.04
- Architecture: amd64
Additional context Add any other context about the problem here.
It should be connected to its network automatically, courtesy of:
https://github.com/spantaleev/matrix-docker-ansible-deploy/blob/da27655ef34999fa924bc0a5e641dbd9ba06f133/group_vars/matrix_servers#L4443C37-L4443C37
