Server Version#: 1.24.1.4931
Player Version#: Issue persists across all clients on the latest client software
I am happy to provide log files if requested.
This issue has actually been a problem for me for a while now (2+ years). It’s just getting noticeable worse. When I click on a movie to watch plex takes a very long time to load the related movies. I have tested this on the web app, the iOS app, and the tvOS app. Seems this is likely more of a server issue than a client issue although some clients take longer to load than others. The tvOS client for example in some cases seemingly just gives up on loading the related movies. This issue may be related to the size of my database. My com.plex app.plugins.library.db is currently 1.05 GB large. In total, I have around 60k entries across my Movie Libraries, and 13,202 episodes in my TV Show library. Is this issue to be expected or are there some optimizations that can be made to the codebase to improve performance?
I have tried servers on both windows and unraid and they both had this issue. Although the load time was worse on my unraid server despite that server having better hardware. All servers have metadata on an SSD. My unraid server uses an NVMe SSD.
I did some testing with the Web Client to see how much better the issue is there and it appears the Web Client is actually just as bad. In a lot of cases, the client just gives up on loading the related movies. As I recall my experience with this issue it’s never been this bad up until very recently. Maybe some changes were made under the covers that were expensive to implement?
In the new Modern UI update, this issue is actually also present in the recommendations page when it tries to load movies to recommend. Can someone help me look into this. This is becoming a real show stopper for me. Would be good it Plex can get ahead of this.
I was able to eventually resolve this issue. Hopefully, this post will help someone else.
TLDR… changing the metadata path in my plex unraid docker container from /mnt/user/appdata/Plex to /mnt/cache/appdata/Plex resolved this issue. Despite the fact that the metadata lives in an NVMe SSD cache pool in both cases putting the data in the cache folder(the default appdata folder) bypasses Unraids FUSE filesystem ensuring the containers read/writes happen directly on the disk significantly improving performance.
I revived my old Windows Server to see how bad the load times were. I was surprised to see it was able to load related content much faster than I remember and much faster than my more powerful Unraid server did. At this point, I wondered if maybe the Plex docker container was not as optimized. After some research, I learned that Unraid does not treat new cache pools the same as it treats its default cache pool which houses the AppData folders for installed docker containers. So I decided to relocate my plex data from the cache drive I created specifically for plex and moved it to the default cache pool Unraid used. Surprisingly this made a world of difference.
