Public Pool Miner Not Registering on Umbrel Node

Open the miner's web UI in a browser. Bitaxe / NerdAxe / NerdQAxe: navigate to the device's LAN IP. Antminer pointed at home: log into the Bitmain dashboard. Find the Pool / Stratum / Settings page and locate the four classic Stratum V1 fields: URL, Port, User (Worker), Password. You're rewriting all four in the next steps.

Set Stratum URL to the LAN IP of your Umbrel node — NOT `umbrel.local`. Find the IP from your router's DHCP client list or from Umbrel's Settings → Advanced. Example: `192.168.0.98`. Do NOT include `stratum+tcp://`, `http://`, `https://`, or any trailing slash. Just the IP. Pin this IP via DHCP reservation in your router so it doesn't move.

Set Stratum Port to `2018`. This is non-negotiable on the default Umbrel Public Pool config. Not `3333` (that's `solo.ckpool.org`'s port and Bitaxe's factory default), not `21496` (that's the hosted `public-pool.io`), not `8332` (that's bitcoind RPC). `2018` is the Public Pool stratum port on Umbrel.

Set Stratum User (Worker) to a real on-chain BTC address. Generate a fresh receive address in Sparrow Wallet, Electrum, Blue Wallet, or a hardware wallet (Coldcard, Jade, Trezor). Must start with `bc1q`, `bc1p`, `3`, or `1`. Must NOT contain `@`, must NOT be a Lightning address, must NOT be an email. Optional worker suffix: one literal dot, lowercase alphanumerics and hyphens only — e.g. `bc1qxyz...abc.bitaxe-garage-01`.

Set Stratum Password to `x`. Public Pool ignores this field for solo address-auth; `x` is the legacy stratum placeholder. Don't put your wallet seed here, don't put a real password, just `x`. Save the miner config and full-power-cycle the device (10 seconds off, then on). Wait 60 seconds. Open `http://<umbrel-ip>:2019` and search for your address — your worker should appear.

If the worker still does not appear after 2 minutes, restart the Public Pool app from the Umbrel App Store: Apps → Public Pool → Restart. Wait for the green Running badge. Recheck the dashboard. If still nothing, escalate to Tier 2 (SSH diagnostics).

SSH into Umbrel: `ssh umbrel@<umbrel-ip>`, password is what you set during onboarding (default `moneyprintergobrrr` if unchanged — change it). Windows users: install OpenSSH client or PuTTY. Mac/Linux: ssh is built in. Then run `sudo docker ps | grep public-pool`. Confirm a container with status `Up X minutes` and ports `0.0.0.0:2018->2018/tcp`. If absent or `Restarting`, the app is not running — reinstall from the App Store.

Read the container logs: `sudo docker logs --tail 200 public-pool_pool_1` (substitute the container name from `docker ps` if different). Look for: `Error getmininginfo Error: ESOCKETTIMEDOUT` (bitcoind RPC unreachable, see Step 10), `ECONNREFUSED` to `:8332` (bitcoind not listening, see Step 10), `EADDRINUSE :::2018` (another process owns port 2018, see Step 11), `initialblockdownload: true` (Bitcoin Core still syncing, see Step 9). The first error line tells you exactly what to fix.

Verify Bitcoin Core is fully synced: `bitcoin-cli getblockchaininfo`. You need `chain: "main"`, `initialblockdownload: false`, and `verificationprogress` very close to 1.0. If `initialblockdownload: true`, the only options are: wait it out (1-7 days on consumer hardware), point Public Pool at an already-synced external node, or use the hosted `public-pool.io:21496` until your local node is ready. Public Pool cannot construct block templates against an unsynced chain.

Confirm RPC reachability between Public Pool and Bitcoin Core. Edit `~/umbrel/app-data/bitcoin/data/bitcoin/bitcoin.conf`: ensure `rpcallowip=172.16.0.0/12` (Public Pool's Docker network) and `rpcbind=0.0.0.0:8332`. If your config has `onlynet=onion`, that's Tor-only mode — Public Pool's clearnet RPC will fail. Either remove `onlynet=onion` (privacy trade-off) or run Public Pool against a separate clearnet bitcoind. Restart Bitcoin Core after any config change: Umbrel App Store → Bitcoin → Restart.

Test the stratum listener from another LAN host (your laptop, NOT the Umbrel itself): `nc -v <umbrel-ip> 2018` on Linux/macOS, or `Test-NetConnection -ComputerName <umbrel-ip> -Port 2018` on Windows PowerShell. Expected: `Connected to <umbrel-ip>`. If `Connection refused`, the listener is not bound externally — restart the app. If `Connection timed out`, a firewall is dropping the packet — check `sudo iptables -L -n | grep 2018` on Umbrel and your router's firewall rules. If `host not found`, the IP is wrong — recheck Step 2.

Verify port 2018 is bound to all interfaces, not just loopback: `sudo ss -tlnp | grep 2018` on the Umbrel. You want `LISTEN 0.0.0.0:2018`. If you see `127.0.0.1:2018`, the container is loopback-only and unreachable from the LAN — restart the Public Pool app to refresh the port mapping. If a different process owns 2018 (rare), stop that process or relocate Public Pool to a different host port via custom Docker compose override.

Pin a static LAN IP for your Umbrel via DHCP reservation in your router. Reserve the Umbrel's MAC address to a specific IP (e.g. `192.168.0.98`). Reboot Umbrel. Confirm the IP didn't change. Update the miner config to the pinned IP. This is the single most-effective preventive against intermittent `umbrel.local` mDNS failures and the resulting `worker disappeared from pool` problem.

If miner and Umbrel are on different VLANs / Wi-Fi networks, open an inter-VLAN ACL for TCP 2018. On Unifi: Settings → Routing & Firewall → Firewall → LAN IN → Add Rule → Allow TCP 2018 from miners-VLAN to services-VLAN on `<umbrel-ip>`. On pfSense / OPNsense: Firewall → Rules → add corresponding pass rule. Test with `nc -v` from a host on the miners VLAN before pointing the Bitaxe — confirm the connection passes the inter-VLAN firewall.

If Tier 1-2 are clean but the worker still doesn't register, run a packet capture on the Umbrel during a miner reboot: `sudo tcpdump -i any -n -w /tmp/pubpool.pcap 'tcp port 2018'` for 60 seconds. Move the pcap to a workstation, open in Wireshark, follow the TCP stream. You should see SYN, SYN-ACK, ACK from the miner's IP, then JSON-RPC `mining.subscribe`, `mining.authorize`, `mining.notify`. If the JSON-RPC never arrives, the miner is being blocked at the application layer — capture the rejection reason from `docker logs` and file on https://github.com/benjamin-wilson/public-pool/issues with both artifacts.

Stop DIY and escalate. Container is running, port is reachable from miner's network, BTC address is valid on-chain, Bitcoin Core is synced, packet capture shows clean stratum on the wire — and the dashboard still shows 0 workers. That is a public-pool application bug, not your config. Open a D-Central support ticket at https://d-central.tech/contact/ with screenshots of: miner Stratum config, `sudo docker ps | grep public-pool` output, last 100 lines of `sudo docker logs public-pool_pool_1`, and `bitcoin-cli getblockchaininfo`. We see this pattern weekly and will triage it inside one ticket.