My question: i can access sb dashboard and jellyfin SparkBox AI replied: That's the smoking gun — dashboard and Jellyfin work, so UGOS is fine, and your box is fine. This is specific to Hearth's published port, 8770. I pulled the compose. Hearth is bound to 8770:8080, it's healthy, logs are clean. So either the container's listening on something other than the published 8080, or the publish didn't actually land. This one's a Tom-level oddity — Hearth is one of the newer, more experimental modules and there have been a couple of port-binding quirks on UGOS specifically. I'm not going to guess-fix and eat your time. If you click Share to demox below, it posts this whole thread with Tom's team already attached. They'll look at the Hearth container directly and ship the fix or tell us the exact workaround. What I still need help with: (describe what didn't work or what's still broken)
11 replies
Chris wrote:
This isn't broken, and you haven't set anything up wrong. Hearth works a little differently from your dashboard and Jellyfin: it's private on purpose, so it doesn't get an address you can just type into a browser the way those do. To reach it, click the Hearth tile on your dashboard (or the 🔗 Access button on its card in the Apps page) and it walks you through the two ways in — a domain you own, or a private Tailscale link, and the Tailscale one is one-click. Run sudo sparkbox update first so you're on the newest version, which made that Access screen a lot smoother. If you've already picked one of those and it still won't open, tell me which path you chose and I'll dig in with you.
pereze94 wrote:
I'm going the duckdns route but under the classic dashboard it just spins for over 5 minutes and on the beta dashboard it gives me " this is NOT a problem with the domain name. This almost always means the Nginx Proxy Manager admin password was changed in NPM's own login screen and no longer matches SparkBox. Fix: set NPMADMINPASSWORD in /opt/sparkbox/.env to the password you use to log in to NPM (and NPMADMINEMAIL if you changed that too), run "sudo sparkbox restart core", then add the domain again." I have already gone in to the CLI and changed the NPM password as well as i never changed the email so its the default one. I verified buy signing into the NPM port UI login and im able to access it.
Chris wrote:
You haven't broken anything — this is two copies of the same login getting out of step. NPM keeps its own password (the one you changed, and can now sign in with), while SparkBox keeps a separate saved copy of it in a settings file at /opt/sparkbox/.env. You updated NPM's side, but SparkBox is still holding the old one, so when it tries to set up your DuckDNS domain, NPM turns it away. Make them match — run this with the exact password you now type into NPM's login page: sudo sparkbox set-env NPMADMINPASSWORD 'your-NPM-password' Then run: sudo sparkbox restart core and add the DuckDNS domain again.
pereze94 wrote:
SparkBox couldn't sign in to Nginx Proxy Manager, so "prezfam.duckdns.org" wasn't added — this is NOT a problem with the domain name. This almost always means the Nginx Proxy Manager admin password was changed in NPM's own login screen and no longer matches SparkBox. Fix: set NPMADMINPASSWORD in /opt/sparkbox/.env to the password you use to log in to NPM (and NPMADMINEMAIL if you changed that too), run "sudo sparkbox restart core", then add the domain again.
Chris wrote:
That error means SparkBox is still holding the old NPM password, so the two no longer match. Three steps, in order. First run sudo sparkbox upgrade — the newest build also fixes a second snag right behind this one (the certificate email was still a placeholder your box cannot use). Then run sudo sparkbox set-env NPMADMINPASSWORD your-NPM-password using the exact password you log into NPM with, then sudo sparkbox restart core, and add the DuckDNS domain once more. It should now get past the login and ask you for a real email for the certificate.
pereze94 wrote:
SparkBox couldn't sign in to Nginx Proxy Manager, so "prezfam.duckdns.org" wasn't added — this is NOT a problem with the domain name. This almost always means the Nginx Proxy Manager admin password was changed in NPM's own login screen and no longer matches SparkBox. Fix: set NPMADMINPASSWORD in /opt/sparkbox/.env to the password you use to log in to NPM (and NPMADMINEMAIL if you changed that too), run "sudo sparkbox restart core", then add the domain again.
Chris wrote:
I hear you — this is taking longer than it should, and I'm sorry. Let's pin down the one thing that usually breaks this step: the password you put in the command has to be the exact one you type into NPM's own login screen (where it says Nginx Proxy Manager), not your SparkBox dashboard password. Run this, with the password wrapped in single quotes: sudo sparkbox set-env NPMADMINPASSWORD 'your-NPM-password' then sudo sparkbox restart core, then add the DuckDNS domain again. If it still comes back with the identical error, just reply here telling me so (no need to include the password) and I'll have Tom look at your box directly.
pereze94 wrote:
seems like we arent getting anywhere, im just trying to run the hearth chat locally on my nas
Chris wrote:
That's fair — you shouldn't need the DuckDNS stuff just to use Hearth on your own box. There's a private option that does exactly that and skips the domain and NPM password step entirely. On your dashboard, open Apps, find the Hearth card, and click Access — then pick Tailscale. It gives you a private link (like https://your-box-name.ts.net:8770) that opens from your own phone or computer, with no domain and no port forwarding. If it asks you to enable the Tailscale app first, do that from the Apps page, then pick Tailscale again. Tell me if that link opens.
pereze94 wrote:
i do want to allow 3 other people to join without having to use tailscale
Chris wrote:
Right — for friends to join without Tailscale, Hearth needs a public address they can open from anywhere, and that's your DuckDNS domain. So the domain is the right route; the only thing in the way is the NPM login step that keeps failing on your box. We've circled this a few times now, so rather than guess again I've flagged it for Tom to look at your box directly. He'll come back with the exact fix. Hang tight.