SOLVED: Container won't deploy after server restart, startup probe failed

[BourbonDoc]

Twice now after restart my Jellyfin container on TrueNAS Scale 24.02 is stuck Deploying after restart. I get the following error:

Startup probe failed: Get "http://172.16.2.46:8096/health": dial tcp 172.16.2.46:8096: connect: connection refused


No issues show up in the log, which you can see here: https://pastebin.com/1H28irtx.

Last time I had to delete and reinstall and everything worked fine, but I'd rather not have to do this every time after restart. All 20 of my other apps work fine. I tried rolling back to the previous Jellyfin as well as current 10.9.11, no change. Any assistance is appreciated!

[TheDreadPirate]

Can you share your configuration for the container?

[BourbonDoc]

How do I do that on Kubernetes? Running TrueNAS Scale and in the container I don't see a good eye readable config file. These are what's in the config folder:
branding.xml encoding.xml logging.default.json metadata.xml migrations.xml network.xml system.xml users

[TheDreadPirate]

Screenshots of the TrueNAS UI is fine.

[BourbonDoc]

Pretty much stock. I've tried it on "Host Network" and not. I usually run it with Host Network off. This config works no problem on a fresh install and had been working fine for about a month before a power outage gave me a new IP from my ISP. Then it happened again when one of my kids unplugged the modem, and now with hurricane Helene. It's definitely something about the new WAN IP, but I have no idea what.

---

Also here's a log: https://pastebin.com/D6AwTPJU

[TheDreadPirate]

Soooo.....what is the problem you are having? The TrueNAS UI says that it is "deploying" constantly? The log looks fine. Is TrueNAS forcing the container to restart because it can't determine its status?

But I noticed that the bridge IP is different than what you mentioned in the first post.

Code:

Startup probe failed: Get "http://172.16.2.46:8096/health": dial tcp 172.16.2.46:8096: connect: connection refused

It appears to be using a different address now. Not sure how big your bridge subnet is.

Code:

2024-09-27 13:57:02.005828-04:00[13:57:02] [INF] [1] Jellyfin.Networking.Manager.NetworkManager: Filtered interface addresses: ["127.0.0.1", "172.16.4.181"]

Also, it looks like you're allowing TrueNAS to manage the container storage. Ideally you would create a dedicated, static space in TrueNAS and use that directory for /config.

[BourbonDoc]

It gets stuck in "Deploying" and the web UI is inaccessible. It is a different IP as its the one assigned in Kubernetes and changes when I redeploy.

I switched over to the TrueNAS beta last night that uses Docker instead of Kubernetes containers. The problem persists. I was super hopeful docker would help. I guess I'm in for another remove and reinstall of Jellyfin.

[TheDreadPirate]

I mentioned in my last post that you weren't providing a static/permanent path for Jellyfin to store data in. That may be part of the problem. Or may be THE problem.

[BourbonDoc]

Tried that and it did deploy. There seems to be sticky data in the truenas ix-application storage locations. Unfortunately now I can't get hardware acceleration to work. TrueNAS official and portainer container docker install both have fatal player failures when I try to use hardware acceleration. Confirmed my graphics card (A380) is showing up in dev directory. It also seems to work fine with plex install.

[TheDreadPirate]

Is it transcoding in Jellyfin but not Plex? What kernel is the beta TrueNAS using? Which image are you using? Our official image or the linuxserver image? What kind of video is it? 4K HDR?