Server Version#: 1.43.2.10687-563d026eaPlayer Version#: 4.159.0
I just subscribed to a monthly plex plan to try it out and I am about done with it. This is NOT what a paid media server is for, and my patience is almost at zero with it. I’ve tried both the stable and beta versions. I will not enable UPnP on my network. I appreciate any help.
My own Plex Media Server keeps disappearing from app[.]plex[.]tv even though Remote Access shows “Fully accessible outside your network” with a green check when I view it on the LAN.
The server is running, fully reachable on LAN, and the local Remote Access UI shows everything healthy — but the server vanishes from the server list on app[.]plex[.]tv after a few hours. It returns after a container restart, then disappears again hours later. Repeatable.
Environment
PMS: 1.43.2.10687-563d026ea (Plex Pass beta)
Web client: 4.159.0
Deployment: Official Plex Docker container, Ubuntu VM on Proxmox
PMS network: Docker bridge mode, ports published to host
Firewall: pfSense; manual port-forward TCP 32400 → server:32400; UPnP/NAT-PMP disabled
DNS: verified no Plex domains blocked
Account: Plex Pass owner; server claimed by my account
Reverse proxy
In addition to the direct port-forward, I have a reverse proxy path set up:
haproxy on pfSense (frontend on 443, terminating a Let’s Encrypt wildcard cert for my domain)
Hostname dedicated to Plex routes to a backend that forwards plain HTTP to PMS:32400 internally
Backend pass-thru sets option forwardfor, option http-server-close, timeout tunnel 1h, and X-Real-IP / X-Forwarded-Proto / X-Forwarded-Port headers
This URL is registered in Plex Settings → Network → “Custom server access URLs”
Verified externally with curl /identity returning the expected XML and machineIdentifier with a valid cert chain
The Local web UI ALWAYS accessible…
My server shows up right away after container restart and everything works fine… for a few hours
Outbound HTTPS to all plex[.]tv API endpoints returns 200
pubsub EventSource connection establishes and stays open
plex[.]tv reachability check workers reach the server, PMS responds 200 to GET /identity
No issues with pub IP detection
Querying plex[.]tv/api/v2/resources directly returns the server with all four connections registered (LAN plex[.]direct, reverse proxy, WAN plex[.]direct, Relay), presence:true, publicAddressMatches:true, natLoopbackSupported:true
Only in the log may be related, I have no idea - the hourly state demotion that doesn’t republish. About every hour, this sequence appears in the log:
CANTUSEURLS??? GET ..plex[.]direct:32400/identity
WARN - NAT: PMP, timed out waiting for response.
Request: [:xxxxx (Allowed Network (Subnet))] GET /identity
… User-Agent => PlexMediaServer/1.43.2.10687
Request came in with unrecognized domain / IP ‘’ in header Host;
treating as non-local
MyPlex: mapping state set to ‘Mapped - Not Published’
HTTP/1.1 (0.1s) 200 response from GET /identity
MyPlex: Last published value didn’t change, we’re done.
PMS makes an outbound GET to its own WAN-IP plex[.]direct hostname. On any pfSense + single-WAN + port-forward setup that connection routes up to the firewall, hits the firewall’s own WAN IP, gets handled by the inbound port-forward, and bounces back to PMS. The inbound request shows up with PMS’s own user agent, confirming it’s the same request. The GET returns 200 — the test succeeds.
At the same time hourly, on a different thread, PMS demotes state to Mapped - Not Published regardless of the GET outcome (the demote and the 200 response happen at the same millisecond on different threads). Then PMS evaluates the new state and logs “Last published value didn’t change, we’re done” and does not send a republish to plex[.]tv. State stays demoted until the next hourly cycle, which does the same thing.
Am I wasting my time here, does plex require UPnP to be enabled? The docs say no.
What is the hourly outbound GET ..plex[.]direct:32400/identity for? Not documented anywhere I can find. Not a published reachability check (those come from plex[.]tv workers with User-Agent: ``plex[.]tv). On networks without NAT loopback it would silently fail.
Why does PMS demote Mapped → Mapped - Not Published on the hourly tick when the self-fetch succeeded? The demote is unconditional and on a different thread from the GET, so it’s not caused by the GET outcome.
Why does the hourly demotion not trigger a republish? “Last published value didn’t change” — but state went from Mapped to Mapped - Not Published. The change isn’t being propagated to plex[.]tv. Intended? What are the states?
Anyone else on a similary version 1.43.2.10687 seeing the same things in the logs? Does it correlate with disappearance of your library from app[.]plex[.]tv?
I haven’t even started troubleshooting the live tv yet, this is not looking good.
