Plex Library is slow/Home Screen Movies and Movies screen fail to load

Server Version#:1.42.0.9975
Player Version#: 4.148.0

Opening New Thread per request

Home screen load fails for movies and does not display, and then when you go to movies it fails. Can click on retry and then after about a min they will come up and display. going back to home screen still does not refresh movies. have to exit and come back in and then they will display. I see this behavior, on Roku, IOS, windows and it is consistent behavior leading me to think it is the server. I have used ChuckPa tool (which is great) on DB, and performed optimize and clean bundles thru PLEX UI. com.plexapp.plugins.library.db is about 1.4G. Plex Version 1.42.0.9975. Running on Windows 11 Pro 13th Gen Intel(R) Core™ i7-1365U with 32G Ram and running on NVMe SSD dedicated for Plex separate from SDD NVMe for OS on different drive.

Plex Media Server Logs_2025-08-08_07-44-04.zip (3.5 MB)

any suggestions I should try to resolve the performance issues? This has been happening for several releases now.

The held transaction for too long warning is frequently seen in the logs. Held transaction for too long (D:\gha\plex-media-server\plex-media-server\Statistics\StatisticsManager.cpp:288): 1.546875 seconds

[Req#8c4cde] SLOW QUERY: It took 906.250000 ms to retrieve 24 items.
[Req#8c4cd3] SLOW QUERY: It took 625.000000 ms to retrieve 24 items.
[Req#8c79c0] SLOW QUERY: It took 406.250000 ms to retrieve 50 items.

ChuckPA DBRepair Log
2025-08-03 07.05.13 – Session start: Host is Windows 10 (Build 26100)
2025-08-03 07.05.13 – Stop - START
2025-08-03 07.05.14 – Stop - PASS
2025-08-03 07.05.14 – Auto - START
2025-08-03 07.10.24 – Repair - Export databases - PASS
2025-08-03 07.12.04 – Repair - Import - PASS
2025-08-03 07.12.41 – Repair - Verify main database - PASS
2025-08-03 07.12.44 – Repair - Verify blobs database - PASS
2025-08-03 07.12.44 – Reindexing Main DB
2025-08-03 07.12.58 – Reindexing Blobs DB
2025-08-03 07.13.04 – Reindexing complete.
2025-08-03 07.13.04 – Moving current DBs to DBTMP and making new databases active
2025-08-03 07.13.04 – Auto - PASS
2025-08-03 07.13.04 – Start - START
2025-08-03 07.13.04 – Start - PASS
2025-08-03 07.13.04 – Exit - Deleted temp files.
2025-08-03 07.13.04 – Session end. 08/03/2025 07:13:04

image

Is there anything I could purge in the DB that would help the performance?

Anyone else having this kind of problem?

Update:

I have given Sqlite 2G for its Cache and installed DymaxIO. WIth these 2 changes, The Home screen and Movie screen are no longer timing out and I am not getting the Something went wrong message.

My assumption is this has helped the DB performance and loaded most things into Cache with Plex for DB and common files.

Assume Plex Support team knows about the DB performance problems with Large DB. I am not that familiar with SQLite. Is there a way to see poor performing queries? Maybe index is missing or needs to be updated.

Open to suggestions.

Bad idea.
You will run out of RAM very fast. This is not a file cache for the DB file.
Put it back to default 20MB, then increase it only in further 20MB steps. Always compare performance before and after. If there is no benefit measurable anymore, stop increasing it.

How large is your primary Plex database file?

The description of this software reads like snake oil.

DB size is 1.4G for primary. my Plex Directory is 160G last time I checked. I was skeptical at first with the DymaxIO. I tried each action individually and did not add dymaxIO until the end of testing of the memory size. Understand it is not a File cache. RAM for PLEX & DB.

I have 32G of Ram on the system. I did start small in my testing and did small increments and went up in 100M increments. More I added the performance improved. still had the timeouts, but they reduced and the held transaction times reduced. I had plex at 1G DB cache and then added DynmaxIO. once I did that, the timeouts stopped.

I bumped up more ram and load speed improved. I had plenty of unused ram, so went with 2G and it seemed to help. PLEX within task manager has been hovering around 1.6G of ram usage and not going up to the 2G.

I can not see from the logs what is happening behind the scene and what is PLEX memory usage vs SQLlite. What I do know is both of these seems to have help minimize the performance issue. Assuming the unknown error is related to timeout error not handled properly. Just a guess.

couple days now, and still seems to be holding up. still a little lag for it to come up on the screen. 20 secs approx, but they come up and do not get the error. Is there something I should do to capture or try some testing to help? Open to try some testing.

I was thinking about bumping up the memory to 64G to see what that does. as I am typing this… PLEX memory consumption dropped to just under 1G since yesterday

I also setup Transcoding directory on a Arsenal Dynamic Ram drive. (just so you know my setup, not related to this issue that I know of)

just re-tested. Approx 15 seconds for home screen to come up. with the 2G cache and Dynmaxio

Just did a test and dropped to the default as you mentioned. it failed to load movies on home screen and when you click on movie tab, it fails with the unknown. The small increment increase did not work for me either. once I had 800mb, it starts to work without having to do many many retries. Movie screen takes approx 22 seconds to come up with my configuration. 2G might be overkill, but I have the memory and appears to not use all of it yet. But is working sluggish, but not erroring out.

something you would like me to test, please let me know how I can help.

I’m afraid you don’t.
The thing is: the cache is for each database session, not globally.
Which means: the more clients are concurrently interacting with your server, the more RAM is consumed, because each client gets its own cache.

With a DB file size of over 1GB, you should definitely check if you are affected by the database bloat issue: Library.db size more than doubled in latest version - #410 by OttoKerner

You’re right, I did not understand that. From the description on the Plex UI, it did not say that.it was a session setting.. I did think it was global setting

I thought ChuckPA’s tool handled the bloat issue, but I will try the method in your post and report back. I hope this is the issue and will help reduce it.

AFAIK, only the Linux version did, and that only for a few short days.
Once it was known that PMS will get the cause fixed, and also get a mitigation for bloated DBs as part of the Scheduled Tasks,
@ChuckPa did remove it from the DB repair script again.

I processed your script, appears my DB had a small number of bloat records so it only reduced 1M. Appreciate your help. Anything else you would recommend I try?

Successfully processed 1 files; Failed processing 0 files
New Size: 1410953216 bytes
Moving existing DB to ‘-bloat’
1 file(s) moved.
Removing WAL and SHM if exist
Making new DB active
1 file(s) moved.
Restoring permissions (check manually if needed)
processed file: com.plexapp.plugins.library.db
Successfully processed 1 files; Failed processing 0 files
New databases directory contents
Volume in drive B is New Volume
Volume Serial Number is AA80-276B

08/12/2025 06:35 PM 1,410,953,216 com.plexapp.plugins.library.db
08/12/2025 06:35 PM 1,411,985,408 com.plexapp.plugins.library.db-bloat

When using Chuck’s script what options are you using?

I am using windows powershell version with Auto

the log above was from OttoKerner’s script from this post

You seem to have an insane amount of media.
Do you have many photos? Or are these all videos?

It is big. Videos, very view photos

I installed the latest beta release. I have not changed any setting, but memory usage for Plex has dropped and hovering around 750M

Performance may be slightly better

update on timing:

I Would need someone to help confirm, but this latest release of Plex something has changed. Home screen loaded in avg of 5 seconds and movie was the same. That is a significant difference. I did not seeing in the release log mention of any performance enhancements.

I will do some more testing to see, but with my current config and this release. seems to have removed most of the sluggishness. it is usable again.

I will try changing setting back down as Otto you mentioned. I can update on what happens.

Are you using any external tools which manipulate the database?
For instance create a high number of collections, or reappropriate genre tags or sharing labels for other uses?

I do not have a large amount of collections or added many tags. I use only 1 tag for sharing/filtering. I do use Kometa. I have not looked thru the python code, but I did think it used the PLEX APIs and not direct DB. but not 100% sure about that. I have been using it for many years.

If there is one piece of software which can cause you grief with a large database, it’s this.

what part of it causes issues? It does have some helpful functionally. are their certain tables I should do some analysis on?