[BUG] cURL request does not work

[NGDM]

Is there an existing issue for this?

• I have searched the existing issues

Current Behavior

I have this docker running through network_mode: 'service:gluetun' together with another container.
cURL requests to this other container via localhost:port used to work, but since 4.6.1-r0-ls297 the requests don't arrive and they don't return.

I believe something changed in the c-ares package which causes this: 4.6.0-r0-ls296...4.6.1-r0-ls297
It seems they changed something to their DNS resolver between 1.21.0 and 1.22.1: https://c-ares.org/changelog.html

Expected Behavior

cURL request arrives and returns.

Steps To Reproduce

1. Run 4.6.0-r0-ls296 in docker with network_mode via container A
2. Run different container B with running webservice on port X (container A has open port X)
3. Open console for qBit-docker and make cURL request via localhost:portX to webservice of container B
4. Request arrives, response received
5. Repeat steps 1-4 with 4.6.1-r0-ls297
6. Request never arrives, no response received, console stays open

Environment

- OS: Synology DSM 7 (Docker version 20.10.3, build 55f0773)
- How docker service was installed: Synology Package Center

CPU architecture

x86-64

Docker creation

qbittorrent:
    image: 'linuxserver/qbittorrent:4.6.1-r0-ls297'
    container_name: qbittorrent
    environment:
      - PGID=$DOCKER_GID
      - PUID=$DOCKER_DOWNLOAD_UID
      - TZ=$TZ
      - WEBUI_PORT=8080
      - UMASK=$UMASK
    volumes:
      - '$DOCKER_PATH/qbittorrent/config:/config:rw'
      - '$DOWNLOADS_PATH:/data/downloads:rw'
    network_mode: 'service:gluetun'
    restart: 'unless-stopped'

Container logs

[migrations] started
[migrations] no migrations found
───────────────────────────────────────

      ██╗     ███████╗██╗ ██████╗ 
      ██║     ██╔════╝██║██╔═══██╗
      ██║     ███████╗██║██║   ██║
      ██║     ╚════██║██║██║   ██║
      ███████╗███████║██║╚██████╔╝
      ╚══════╝╚══════╝╚═╝ ╚═════╝ 

   Brought to you by linuxserver.io
───────────────────────────────────────

To support LSIO projects visit:
https://www.linuxserver.io/donate/

───────────────────────────────────────
GID/UID
───────────────────────────────────────

User UID:    1031
User GID:    65539
───────────────────────────────────────

[custom-init] No custom files found, skipping...
WebUI will be started shortly after internal preparations. Please wait...
Connection to localhost (127.0.0.1) 8080 port [tcp/http-alt] succeeded!
[ls.io-init] done.

[github-actions]

Thanks for opening your first issue here! Be sure to follow the relevant issue templates, or risk having this issue marked as invalid.

[drizuid]

We do not offer support for tunneling a container through another.

[Outsmart2383]

@NGDM did you find a solution for this? Thanks!

[NGDM]

@NGDM did you find a solution for this? Thanks!

It seems this got fixed along the way with a new c-ares version.
Not sure when, but I'm on latest version now and am able to make local requests again from within the container.

[d-e-x-x]

I don't have a fix yet, but I can offer a workaround. This seems to be related to the default k8s /etc/resolv.conf. The nameserver is set to gluetun's unbound, and something there isn't working correctly with the search and options config lines from k8s. I manually commented those lines out in this container and requests started working.

Breadcrumb: qdm12/gluetun#281 (comment)