(Win 10 WSL Ub 24.04) Current state: Launching the webui, I get a Jellyfin Startup page saying "Jellyfin Server still starting. Please wait", a bit of the log, and a link to the log. Beginning of the log I have [2026-06-12 00:00:17.783 -04:00] [WRN] [82] Microsoft.Extensions.Diagnostics.HealthChecks.DefaultHealthCheckService: Health check "StartupCheck" with status Degraded completed after 0.0002ms with message '"Server is still starting up."' ....repeated a zillion times. All before I looked at it this morning. Then some entries from what I assume was attempted update/upgrade/restart/etc. All with "Server is still starting up" repeated after each. (full log in next post if it will fit) All indicators show the container as "healthy", but running test-media.sh, I get ! WARN jellyfin-media — http://localhost:8096 returned HTTP 503 :8096/health says "Degraded" History: Things were going fine for a while, even the moonfin/seer link finally worked. Then it didn't. Then I saw an update that said it fixed it. I tried an "update", which told me to "upgrade". I've since tried various re-install/upgrade/update and I appear to be on the current version (1.6.236 at the time of this post). In the middle of this shotgun-mess of troubleshooting, I also updated Docker, which messed with the WSL localhost a bit, but I fixed that so I can access everything via 192 etc. I've run through a number of things I saw in the docs but now I've been going in circles and not sure which end is up anymore. Anyway, out of time for at least today, maybe the weekend. Am I at a "scrub and start over" situation?
6 replies
fellini wrote:
Log is 628k chars so I guess not. Nothing particularly obvious that I can see.
tomspark wrote:
You didn't break anything — the container's healthy, it's Jellyfin itself stuck mid-startup, which is almost always it choking on its own database right after a version change (and WSL makes that first post-update start extra slow). Two steps. First, restart just Jellyfin and give it a full 5 minutes before judging — on WSL that initial run can really crawl: sudo docker restart sb-jellyfin-media If it's still 'starting up' after that, the useful clue isn't those repeated StartupCheck lines — it's the LAST 30 or so lines of the log, where the real error hides. Paste those and I'll pin it down. Don't delete anything yet.
fellini wrote:
So it turns out I had a bigger Windows problem than I thought. Decided I'd finally make the effort of scrubbing this old box and putting Mint on it. In the process of rebuilding everything, restoring my media backups now. More to come.
tomspark wrote:
Glad you got to the bottom of the bigger Windows problem — and honestly Mint is a great base for running this stuff, you'll probably have a smoother time of it. No rush at all; get your media restored and set back up at your own pace. If Jellyfin ever hangs on that 'still starting' screen again once you're back up, give me a shout — that exact one's on my radar. Good luck with the rebuild!
fellini wrote:
Even bigger - the old nas drives in my pool were failing. Probably because they were actively being used for once this past month or so. Anyway, I need to totally re-think my mediaserver setup since I should probably finally retire this old 3rd-gen. Feedback on "smoother", though. I was running in to a number of spots between your docs and your scripts where case-sensitivity (TV vs tv, etc) wound up making multiple redundant dirs. It even tried to put them on root /data at one point. How much of this was crossed with permissions because of drive issues is unclear. But there's a fair bit of contradiction that could use some cleanup. A simpler straightforward non-NAS "linux with a big media drive that has a folder structure already set up on it" guide/option might help. Or at least get rid of the old legacy info. That's all for now. Thanks for the help and the learning experience, and I'll be back whenever I figure out where I'll put this thing next. Cheers!
tomspark wrote:
Sorry about the drives — rough way to find out, but better now than after you'd piled more onto them. And thank you for the case-sensitivity note; that's genuinely useful. 'TV' vs 'tv' turning into duplicate folders (or worse, landing at /data) is exactly the kind of contradiction we want gone, so I'm checking where our docs and scripts disagree on folder names. The 'plain Linux box with a big media drive that's already organized' guide is a good shout too — I've written it down. Whenever you land on your next setup, come back and I'll help you get it dialed in. Cheers, and thanks for the detailed feedback.