(SQLite?) When server scans library clients unusable

Server Version#: 1.28.2.6151
Player Version#: Any

Server:
Ryzen 7 5800X
32GB RAM Available (after RAM disk for scratch is allocated)
Dedicated EVO 970 Plus 2TB for metadata
40Gb Ethernet connection to storage server
1Gb Up/Down internet connection for clients

When my server does a library scan initiated manually or by schedule the clients become unusable, whatever is playing craps out and browsing the library is wickedly slow and going into a show or movie doesn’t show anything and sometimes displays an error.

I have had 8 clients connected to the server without any issues however a library scan renders it unusable for even one client. I have tested from the server and while it is scanning browsing the library is unusable.

During scanning the CPU averages a few % with short peaks of 20%, RAM usage is < 10GB, disk average active time < 1%.

I don’t know what to do to resolve this, any ideas? Thanks.

Error that shows up sometimes when browsing library while a scan is happening.
Browse error when scanning

User _plx_svr seems to have had the same issue {plex-scanner-drops-plex-media-server-cpu-usage-to-0-and-kicks-clients-off}.

Looking at the logs it seems that after the scanner is complete is when the client playback and browsing issues occur. I’m guessing it’s an issue with the SQLite db.

15:56:24 … completed 100.0% - Scanning Movies

~15:56:35 is when I skipped back 10 seconds and the client errored out and after that I tried browsing shows on the server and it gave me the “Something went wrong” error message.

After about another minute everything went back to working normally.

For these logs I stopped the plex server, restarted it, waited a minute or so for the web interface to become responsive and then I manually started a scan of Movies.

I hope this is not expected behavior, is anybody else having this issue?
Plex Media Server Logs_2022-09-06_15-58-21.zip (511.9 KB)

I also did a “PRAGMA integrity_check;” and the result was “ok”.

After extensive changes in the DB, the server will perform a “vaccum” of the DB file. This operation should only take a short time. And yes, during this operation, the server may not respond to client requests.

Good to know. I have the same behaviour I described even when there are no charges to the DB,. If I run the scan for a second time a couple minutes after its complete it will become unusable again, same as before.

I’m told that server version 1.29.1 could bring a fix which might be related to this.

Thanks for the information and for your time. Do you happen to know if that update is being published soon? As you mention fix, is there an acknowledged bug report for this issue?

I am not able to tell release dates.
The bug tracker is not publicly accessible.

I just updated to version 1.29.0.6244 as it noted under bug fixes:

  • (Butler) The server could become unresponsive during database optimization (#13820)

The issues is significantly worse now.

Updated to the 64 bit version, Version 1.29.1.6260, issue is still there.

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.