My Plex server stopped connecting properly from the TV. I started by checking whether the pod was crashing. Kubernetes showed it Running and Ready, with zero restarts, while the logs kept repeating No user exists for uid 1000.
The main server could answer its identity endpoint. Plex Relay was failing separately, exiting with code 255. Running the relay executable directly reproduced the missing-user error. whoami could not resolve the container's numeric UID either.
My chart ran the container as UID/GID 1000, but its account files had no matching user. LinuxServer documents the image's user/group settings and non-root operation; the mismatch was in my running configuration.1
I added an init container to prepare account-file overlays containing the missing user and group, then mounted those files into Plex. This kept the existing data ownership at 1000:1000.
| Check | Before | After the account fix |
|---|---|---|
| Pod status | Running and Ready | Running and Ready |
| User lookup | UID 1000 had no name | whoami returned plex |
| Sampled relay failures | Repeated exit 255 | No exit-255 failures in the checked log sample |
The later logs showed playback progressing, but also connection resets. That is enough to verify the user lookup changed, and too little to call sustained TV playback fixed.
I also merged a Plex version bump. Its attempted rollout hit an unreachable Kubernetes API, so the live account-fix checks still describe the older image. Confirming the upgrade and a sustained playback session remain open. The useful extra health check here is whether the relay can resolve its own user, alongside the server's HTTP response.
Footnotes
LinuxServer Plex image documentation, including user/group identifiers and non-root operation. ↩