SOLVED: Completely unresponsive when scanning libraries

[gaming09]

trying this next
Edit, tried this, added my g/pg movie library about 280ish movies and same thing it freezes on a completely vanilla install

No plugins on my side, it has to be something locking up the main thread till the scan completes.  It's very easy to replicate, just start up a fresh docker image (official with no additional plugins), point it at a media directory with about 200 movies in.  Have a second tab open with the web UI in a browser, as soon as you start a scan, hit refresh on the web UI and see that it doesn't respond.

I have a pretty big collection and it only takes about 30seconds to scan new media for my movies.  But for my TV shows its a little longer, maybe due to having more folders to traverse, so its not a major issue for me.  However on a new install with a decent sized library it can take a few minutes before the UI is in a usable state.

[gaming09]

new update, ive changed some sonarr/radarr settings its made it about 4% better. for connect ive changed it to on import/ on series delete. so less scans but when it does scan its still freezing

[TheDreadPirate]

When Jellyfin hangs/becomes unresponsive, are there a boat load of ffmpeg processes running?

[gaming09]

on my prod instance i have jellyscrub / introskip run
But i had 2 different tests, 1 disabling all plugins and then 2 a vanilla install and the same issue

When Jellyfin hangs/becomes unresponsive, are there a boat load of ffmpeg processes running?

[toytown]

When Jellyfin hangs/becomes unresponsive, are there a boat load of ffmpeg processes running?

No. "jellyfin" process shows 100% cpu (1 core) and no ffmpeg (or any other processes) are launched.   I have all thumbnail extractions etc turned off.  The UI is unresponsive until cpu goes back down to 0%

[marshalleq]

Yeah, I've been trying out Jellyfin last couple of days and I've just noticed this too. DB is running on Mirrored SSD's, other systems don't have this issue. This has got me wondering if there's other database options I could shift to - MySQL etc. Anyway, I'm all vanilla, brand new install, no added plugins.

Sorry just saw the SOLVED prefix on this one, but there doesn't seem to be a solution, anyone know what solved it?

[TheDreadPirate]

Unsolved. Nothing logged that would explain the behavior. A theory provided by another mod/dev is that some internal OS queue is being maxed out. Caused by slow I/O. A slow CPU, kernel bug? Hard to say.

[Venson]

if its a single core and no "disconcerning" log entries are present, my default guess is the sqlite db is clogged up with requests.