Move Omarchy's Docker bridge off 172.17.0.0/16 when the LAN uses it
The cause and fix hold. Confirmed locally on 4.0.4-1: daemon.json has bip 172.17.0.1/16 and dns 172.17.0.1, /etc/systemd/resolved.conf.d/20-docker-dns.conf has DNSStubListenerExtra=172.17.0.1 (both omarchy-settings backup files), /etc/ufw/user.rules carries the two allow-docker-dns rules, and a grep of /usr/share/omarchy finds no other 172.17 reference, so four places is right. omarchy-dns accepts Custom. #11757 is open and its body does say bip was 172.30.0.1/16 on the reporter's 4.0.3, which quattro and 4.0.4-1 contradict, as the record says. The symptom is wrong in one detail: because docker0 owns 172.17.0.1, ip route get 172.17.0.1 on this machine prints local 172.17.0.1 dev lo, while another address such as 172.17.5.10 prints dev docker0. Not exercised: moving the bridge. Second audit confirmed the corrected text: Re-checked the corrected text on this workstation. daemon.json has bip 172.17.0.1/16 and dns 172.17.0.1, /etc/systemd/resolved.conf.d/20-docker-dns.conf has DNSStubListenerExtra=172.17.0.1, both are listed as omarchy-settings backup files, and quattro install/config/firewall.sh adds the two allow-docker-dns rules. docker.service is enabled. ip route get 172.17.5.10 prints dev docker0 and ip route get 172.17.0.1 prints local 172.17.0.1 dev lo, exactly as the symptom now says. ufw's rule match in /usr/lib/python3.14/site-packages/ufw/common.py treats a rule that differs only in comment as a match, so the delete commands without the comment work. The ufw-docker after.rules use 172.16.0.0/12, which still contains 172.31.0.0/16. Issue #11757 is open and supports the cause, including the reporter's 172.30.0.1/16 bip that quattro and 4.0.4-1 contradict. Not exercised: moving the bridge.
Restarting docker.service stops every running container. Both edited files are pacman backup files of omarchy-settings, so your edits survive updates, but a future omarchy-settings change to them arrives as a .pacnew that must be merged by hand rather than copied over.
Symptom
Hosts on my network that live in 172.17.x.x (the router admin page, a NAS, a VPN subnet) are unreachable from this laptop, while other machines reach them fine. ip route get 172.17.5.10 (any LAN host in that range) answers dev docker0, and ip route get 172.17.0.1 answers local 172.17.0.1 dev lo, because this machine owns that address. Setting the router as DNS with omarchy dns Custom and 172.17.0.1 silently falls back to FallbackDNS, or journalctl -u systemd-resolved shows Configured DNS server loops back to us.
Cause
Omarchy pins Docker's default bridge to 172.17.0.1/16 in /etc/docker/daemon.json ("bip": "172.17.0.1/16", "dns": ["172.17.0.1"]), binds a systemd-resolved stub listener there through /etc/systemd/resolved.conf.d/20-docker-dns.conf (DNSStubListenerExtra=172.17.0.1), and its installer adds ufw rules allowing 53/udp to 172.17.0.1. docker.service is enabled, so docker0 claims the whole 172.17.0.0/16 route from boot, and anything aimed at 172.17.0.1 is answered by the host's own resolved stub. Reported as omacom/omarchy#11757. That report says 4.0.3 had already moved bip to 172.30.0.1/16, but the quattro branch and omarchy-settings 4.0.4-1 both ship 172.17.0.1/16.
Fix
See what the LAN and any VPN use, and pick a /16 that nothing uses. The example uses 172.31.0.1/16:
ip -4 route
cat /etc/docker/daemon.json
**Omarchy 4:** the bridge address appears in four places and all four have to move together, or container DNS breaks.
- Edit
/etc/docker/daemon.json, keeping the other keys. The pool line keeps compose networks off 172.17 as well:
{
"log-driver": "json-file",
"log-opts": { "max-size": "10m", "max-file": "5" },
"dns": ["172.31.0.1"],
"bip": "172.31.0.1/16",
"default-address-pools": [
{ "base": "172.18.0.0/15", "size": 24 }
]
}
- Move the resolved stub listener:
sudo sed -i 's/^DNSStubListenerExtra=172.17.0.1$/DNSStubListenerExtra=172.31.0.1/' /etc/systemd/resolved.conf.d/20-docker-dns.conf
- Move the firewall rules:
sudo ufw delete allow in proto udp from 172.16.0.0/12 to 172.17.0.1 port 53
sudo ufw delete allow in proto udp from 192.168.0.0/16 to 172.17.0.1 port 53
sudo ufw allow in proto udp from 172.16.0.0/12 to 172.31.0.1 port 53 comment 'allow-docker-dns'
sudo ufw allow in proto udp from 192.168.0.0/16 to 172.31.0.1 port 53 comment 'allow-docker-dns'
- Restart, Docker first so docker0 carries the new address:
python3 -m json.tool /etc/docker/daemon.json
sudo systemctl restart docker.service
sudo systemctl restart systemd-resolved
Recreate compose projects (sudo docker compose down && sudo docker compose up -d) so their networks move off the old range.
**Plain Arch:** only bip matters, because nothing else is tied to it. Set "bip": "172.31.0.1/16" in /etc/docker/daemon.json and restart docker.service.
Verify
ip -4 addr show docker0 shows 172.31.0.1/16, ip route get <a LAN host in 172.17> goes out the LAN interface, sudo ss -ulpn | grep 172.31.0.1:53 shows systemd-resolved, and sudo docker run --rm alpine nslookup archlinux.org resolves.