Potential permissions issue with plexmediaserver

Server Version#: 1.40.5.8921
Player Version#: 1.40.5.8921
<If providing server logs please do NOT turn on verbose logging, only debug logging should be enabled>

I completed a new install for Alma Linux 9.6 and installed plexmediaserver-1.42.1.10060-4e8b05daf.x86_64.rpm. Plex started and metadata was displayed correctly. However trying to start a video results in the gold beachball spinning and it eventually times out. I tried an install of plexmediaserver-1.40.5.8921-836b34c27.x86_64.rpm.

Same results. I read a forum page on permissons. I used usermod -a -G markb plex. Which should add plex to the user group markb. Playing a video will now work 1 or 2 times then it reverts back to the spinning beach ball. All the plexmediaserver files are assigned to plex:plex.

All the media files are assigned to markb:markb. The previous installation on Alma Linux 8 had exactly the same permissions and worked fine for months. I’m missing some small (large?) detail. see attached logs files.

I get the same spinning beachball effect on chrome, firefox, iPad Plex app, Windows Plex app.

Plex Media Server Logs_2025-11-08_18-13-20.zip (749.0 KB)

What are you doing here ?

Nov 08, 2025 16:11:31.997 [139943648729912] WARN - Warning: ex: boost::filesystem::last_write_time: No such file or directory [system:2]: “/media/virtual/plexmediaserver/Library/Application Support/Plex Media Server/Plug-ins”, couldn’t check file: “/media/virtual/plexmediaserver/Library/Application Support/Plex Media Server/Plug-ins”

Did you create a systemd override.conf or a symlink?

PMS itself isn’t running right. It crashed several times in these logs.

Yes to systemd override.conf

Here is the override file I have been using.

/etc/systemd/system/plexmediaserver.service.d/override.conf

override.zip (405 Bytes)

please explain the pathname construction because it’s extremely suspicious:

  1. /media - Using the Nautilus-managed /media directory ? not good if the automounter is active
  2. virtual - Is this a network mount / bind / VM of some kind?

What’s the driving force to move from /var/lib and not use /home/plexmediaserver ?

virtual is a virtual directory for mergerfs. I have 14 or so drives merged into a single directory

Automounter is not being used. All drives are mounted in fstab using UUID.

I use snapraid to backup all the media and plexmediaserver metadata and databases.

Before starting PMS the first time, let’s check the obvious (sorry)

  1. Created the base mount point dir (mkdir -p),
  2. Set plex:plex as owner
  3. set permisions chmod -R 755 mount_point_dir
  4. Now do the mount
  5. chmod again
  6. Now let PMS start up the first time and create all the sub structure as well as download all the required submodules (plug-ins – which you don’t have ---- and – likely don’t have the codecs yet either ?)

plexmediaserver permissions

drwxr-xr-x plex plex plexmediaserver

plexmediaserver directory was a working installation on previous OS. I’m supposed to delete some of the previous directories?

You are transplanting PMS from another instance to this host?

How are you doing it ? tar ball ?

no transplant. I had a working AL (AlmaLinux) 8.6 install with a working Plex 1.42 install. There were some issues with the AL 8.6 install. So I wiped the 8.6 install and upgraded to AL 9.6.

Installed plex 1.42 to the new AL 9.6 install and used the existing /media/virtual/plexmediaserver/* files. New OS and plex using the existing working plexmediaserver directory.

UID, GID match and groups memberships recreated ?

Standard procedure when transplanting is a sudo chown -R plex:plex /var/lib/plexmediaserver (reference)

Your problem does seem to mimmick ownership problems because it can’t see any of the Plug-ins directories

Might it be the mergerfs volume creation?

I had some errors in fstab and fixed those. I upgraded to 1.42 but now core-dump?

it is not starting at all. see attached journalctl file.

journalctl.zip (20.1 KB)

How do I get log files when plexmediaserver isn’t running?

I just looked at the databases and both were 0 in size. Restoring them from backups.

I also found a crash dump. see attached.

crash_dump.zip (1.0 KB)

@markbz

Get into the terminal (cmd line)

cd over to where the PMS Logs directory is.

Make a tar.gz of that entire Logs directory.

Attach that tar.gz here

This is Linux telling us it cannot complete the file io.

Nov 09, 2025 17:06:47.498 [139840940841616] ERROR - SQLITE3:0x80000001, 5386, os_unix.c:40311: (19) mmap(/media/virtual/plexmediaserver/Library/Application Support/Plex Media Server/Plug-in Support/Databases/com.plexapp.plugins.library.db-shm) - No such device
Nov 09, 2025 17:06:47.498 [139840940841616] ERROR - SQLITE3:0x80000001, 5386, disk I/O error in "PRAGMA cache_size=512"
Nov 09, 2025 17:06:47.498 [139840940841616] ERROR - SQLITE3:0x80000001, 5386, os_unix.c:40311: (19) mmap(/media/virtual/plexmediaserver/Library/Application Support/Plex Media Server/Plug-in Support/Databases/com.plexapp.plugins.library.db-shm) - No such device
Nov 09, 2025 17:06:47.498 [139840940841616] ERROR - Promise destroyed without handling exception: Promise fell out of scope without resolving or rejecting

I seriously question whether or not this distro & version is compatible with PMS.

See attached

Logs.tar.gz (145.9 KB)

Thanks.

That is what I expected (as I posted above)

Nov 09, 2025 17:06:47.497 [139840940841616] DEBUG - Opening 20 database sessions to library (com.plexapp.plugins.library), SQLite 3.39.4, threadsafe=1
Nov 09, 2025 17:06:47.498 [139840940841616] ERROR - SQLITE3:0x80000001, 5386, os_unix.c:40311: (19) mmap(/media/virtual/plexmediaserver/Library/Application Support/Plex Media Server/Plug-in Support/Databases/com.plexapp.plugins.library.db-shm) - No such device
Nov 09, 2025 17:06:47.498 [139840940841616] ERROR - SQLITE3:0x80000001, 5386, disk I/O error in "PRAGMA cache_size=512"
Nov 09, 2025 17:06:47.498 [139840940841616] ERROR - SQLITE3:0x80000001, 5386, os_unix.c:40311: (19) mmap(/media/virtual/plexmediaserver/Library/Application Support/Plex Media Server/Plug-in Support/Databases/com.plexapp.plugins.library.db-shm) - No such device
Nov 09, 2025 17:06:47.498 [139840940841616] ERROR - Promise destroyed without handling exception: Promise fell out of scope without resolving or rejecting
Nov 09, 2025 17:06:47.591 [139840940841616] DEBUG - [JobRunner] Job running: /usr/lib/plexmediaserver/CrashUploader "--directory=/media/virtual/plexmediaserver/Library/Application Support/Plex Media Server/Crash Reports/1.42.1.10060-4e8b05daf" --version=1.42.1.10060-4e8b05daf --platform=Linux "--platformVersion=9.6 (Sage Margay)" --serverUuid=072faa0a26ef3af070f54fac919c759620773e2a --userId=markb369@yahoo.com --sentryUrl=https://o17675.ingest.sentry.io/api/1233455/ --sentryKey=xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx --vendor=AlmaLinux --model=x86_64 "--device=ASUSTeK COMPUTER INC. ROG CROSSHAIR VIII HERO (WI-FI)" --sessionStatus=crashed --sessionStart=1762729607 --sessionDuration=0

The OS is not constructing and then sustaining a usable file system.

Regrettably , It’s either the MergerFS volume or the Distro version itself.

ouch. Alma Linux 9.6 based off of RHEL 9.

I’ve used RockyLinux successfully .

It seems to be a good successor to Centos.

9.6 is the Alma Linux version. The kernel version is kernel.x86_64 5.14.0-570.12.1.el9_6

Please confirm this is indeed kernel version 9?

Nov 09, 2025 17:06:47.494 [139840853101368] INFO - Linux version: 9.6 (Sage Margay), language: en-US

The current mainstream Linux kernel is 6.x

check /etc/os-release

EDIT: Quick checked. You want to minimally be 5.15 kernel. NAS vendors have all upgraded to that minimum

uname -r

5.14.0-570.60.1.el9_6.x86_64

Ok.. Definitely time to get to a newer kernel. (5.15 or better)

PMS runs on NAS boxes with 5.15 kernels 100% solid