Hmm, I was able to get this to work first try. Can you refresh the metadata for just that one episode then find the Plex Media Scanner Matcher.log file and post that.
Oh awesome! Wasn’t expecting anyone to do that yet but appreciate that you did.
Very good to know!
Hmm, I’m still seeing it here. Is this affecting the library that you upgraded or just all of them in general?
I’ve noticed more and more that TVDB doesn’t have very good coverage for ratings and will fall back to the next source in the list if we don’t find the preferred rating (in the order they’re listed in the preference). We’ve never had ratings for season before, but that might change in the future if we have a good data source for it.
@chrisallen could you add this to that client testing matrix?
We’ll be updating this at some point soon, we’re currently relying on IMDB for this but TMDB has added new data for this which is better that we’ll be moving to.
Once the episode ordering options are in this should be easier to handle.
I only have 1 TV Show library which is the one I upgraded and once refresh completed on deck was gone from the home page. I restored my server (I backup /var/lib/plexmediaserver before doing any tests) and after I restored on deck was back as I would expect.
I can try again tomorrow and see if it happens again. Won’t get a chance to do it again today as I’m trying to limit the impact on my server for those using it.
Do when know if/when the new agent will sort alternative ordering from TVDB? they have data for alternative ordering for a lot of shows going off Arc’s or DVD release etc… which could really help with managing a lot of anime some animes only show as 1-3 seasons but are actually like 6-8 - Hunter x Hunter and Black Clover come to mind.
For tv series/episode parsing the best option is to have ID embedded across entire file structure?
Also, which ID, {tvdb-id} {tmdb-id}, is preferential for the new system? If you have tvdb, will it use that db as source of truth and fall back on tmdb as needed? or vice-versa
Is it worth having both IDs embedded or will that mess up the system?
From what I understand, one ID on the show level will do. There’s no need to repeat the show ID on every episode file.
The IDs are solely used for matching. They don’t influence how the agent is downloading your metadata (or from where).
Not sure if including multiple IDs will do any harm… doing so will certainly not provide any value-add 
here you go… seems not to find the local extra (they’re however still visible on the season level).
Feb 04, 2021 17:46:00.430 [0x700007979000] INFO - Plex Media Scanner v1.21.3.3983-a090db7d3 - Apple Mac x86_64 - build: darwin-x86_64 - GMT 01:00
Feb 04, 2021 17:46:00.430 [0x700007979000] INFO - MacOSX version: 10.16, language: de-DE
Feb 04, 2021 17:46:00.430 [0x700007979000] INFO - Processor Intel(R) Core(TM) i7-7700K CPU @ 4.20GHz
Feb 04, 2021 17:46:00.430 [0x700007979000] INFO - /Applications/Plex Media Server.app/Contents/MacOS/Plex Media Scanner --match --type 2 --log-file-suffix Matcher --item 47238
Feb 04, 2021 17:46:00.433 [0x10e0f5e00] DEBUG - Opening 20 database sessions to library (com.plexapp.plugins.library), SQLite 3.26.0, threadsafe=1
Replacing the special character is however fixing it for me (ū → u).
If it used to work for you… could this be based on how different platforms handle string (file name) comparisons? → as per the log I’m on macOS 11.2
Feb 04, 2021 18:00:34.949 [0x700000abe000] INFO - Plex Media Scanner v1.21.3.3983-a090db7d3 - Apple Mac x86_64 - build: darwin-x86_64 - GMT 01:00
Feb 04, 2021 18:00:34.949 [0x700000abe000] INFO - MacOSX version: 10.16, language: de-DE
Feb 04, 2021 18:00:34.949 [0x700000abe000] INFO - Processor Intel(R) Core(TM) i7-7700K CPU @ 4.20GHz
Feb 04, 2021 18:00:34.949 [0x700000abe000] INFO - /Applications/Plex Media Server.app/Contents/MacOS/Plex Media Scanner --match --type 2 --log-file-suffix Matcher --item 48196
Feb 04, 2021 18:00:34.953 [0x111416e00] DEBUG - Opening 20 database sessions to library (com.plexapp.plugins.library), SQLite 3.26.0, threadsafe=1
Feb 04, 2021 18:00:35.245 [0x111416e00] DEBUG - Scanner: found local extra "/Volumes/video/tv-shows (en)/Naruto Shippūden (2007)/Season 01/Naruto Shippuden (2007) - S01E01-featurette.mkv"
Feb 04, 2021 18:00:35.425 [0x111416e00] DEBUG - Matcher: no sidecar subtitle files found for /Volumes/video/tv-shows (en)/Naruto Shippūden (2007)/Season 01/Naruto Shippuden (2007) - S01E01.mkv
Feb 04, 2021 18:00:35.425 [0x111416e00] DEBUG - Matcher: no valid subtitles found, removing any existing media bundle file
Feb 04, 2021 18:00:35.425 [0x111416e00] DEBUG - We're going to try to auto-select an audio stream for account 1.
Feb 04, 2021 18:00:35.425 [0x111416e00] DEBUG - Selecting best audio stream for part ID -1 (autoselect: 1 language: en)
Feb 04, 2021 18:00:35.425 [0x111416e00] DEBUG - We're going to try to auto-select a subtitle.
Feb 04, 2021 18:00:35.425 [0x111416e00] DEBUG - Audio Stream: -1, Subtitle Stream: -1
Yes, I suspect that’s exactly what’s happening. I may not have time to look at it this week but I’ll get this fixed.
So a TVDB or TMDB id would be synonymous to the system, it just confirms identify and does it’s own backend metadata stuff. Is there a recommended id to insert? My current library uses TVDB agent, but there are a few I’ve either fixed match or had to match using TMDB.
I suppose we’ll only know for sure once the new agent is finalized.
My guess is, it won’t make a difference.
Devs have commented above that the new agent will not source its metadata from a single source. Some other users have already raised questions how this will play with shows that have different episode ids for the same episode (users gave an example from a recent Doctor Who special that’s handled differently on TMDb and TVDb).
Yes those cases are always hard. How do you decide who is truth.
I am asking questions because I will try the beta, but want to setup the library with embedded IDs in the file structure. Just to give it the opportunity to trust that is the series. Then see what it does.
Was this fixed with release: 1.21.3.3983-a090db7d3
@drzoidberg33 Will reading from nfo files be something that will happen with this new scanner or am I wasting my time tracking this post for updates?
See above… – however not 100% sure if this refers to a wider/general support of nfs-based local metadata or reading an IMDB id from an nfs-file (same as the movie agent does).
I’ve been playing a bit with updating some playground tv-show libraries to the new agent (using the library context menu ... > Manage Library > Upgrade Matching).
Works pretty smooth. The only “thing” I noticed is that after the update the option to prefer local/embedded metadata is automatically checked. Not sure that’s the intended default behavior.
It’s not super high priority right now and has some nuances that make it trickier than our support for movies but it’s still on the radar.
This is based off the order of the Local Media Agent you have set up for the legacy agent when you do the upgrade. If the Local Media Agent is on top then we set both those options otherwise just “Use local assets” is checked.
The “Upgrade Matching” option is also only available to Ninjas like you @tom80H - so others won’t see this (but can still upgrade manually by changing the agent in Advanced).
It wasn’t a server update, it’s a fix that was deployed on our backend services so no update from your end required.
In that case it works as designed 
I love the new Agent.
Here is some feedback of just a few things that didn’t match properly in my 200+ TV Show library.
I did manually refresh metadata on those shows with no change.
America’s Got Talent and America’s Got Talent: The Champions were bundled together under America’s Got Talent. It did offer me a Split Apart option, though. However even after splitting, refreshing metadata or fixing match still didn’t bring up America’s Got Talent: The Champions. When I renamed the folder to add the tvdb id in brackets and then unmatched it, then it renamed it but didn’t load new posters. When I refreshed metadata on it, it flipped back to join the other show. I will try Plex dancing with this a bit more and see.
The Alientist was named The Alienist: Angel of Darkness, which is supposed to be the name of the second season only.
The Flash, it put some seasons under The Flash 1990 and some under The Flash 2014. This was my bad due to naming it The Flash and not The Flash (2014). Renaming the folder resolved it.
All in all, very happy with the new agent.