PMS 1.18.x not reachable after upgrade with 100% cpu useage

Plex Media Server Desktops & Laptops
,
1

Server Version#:1.18.5.2309-f5213a238
Player Version#:various
Logs.zip (8.2 MB)
Hello everyone,
since version 1.18.0 I have had problems upgrading my PMS running on a hyper-v win10pro instance. Whenever I upgrade (either via GUI or using the installer manually) I cannot access any library nor the server settings.
When accessing the task-manager I can see that the plex process is consuming 100% cpu ressources. When closing all browser sessions to my server I can see the process is returning to normal values, however as soon as I return it bursts up again.
I have tried accessing the server directly on the VM via 127.0.0.1:32400/web directly, via a different computer and via app.plex.tv. All results are the same.
I can see the interface, but all the on-deck items, settings and libaries won’t load. When i terminate the plex process a regular chrome not-found page is displayed. It seems the UI is still served. Of course I have researched before posting but none of the other thread’s solutions worked for me.
I have tried:

  • repairing / rebuilding via instructions from : Repair a Corrupt Database 1.22.0 and earlier | Plex Support before and after the upgrade.
  • otimizing the database / cleaning the trash etc before upgrade.
  • replacing the database with backed up database after the upgrade.
  • disabled butler tasks and library auto updates / disabled various settings like permitted networks etc…
  • deleting all music libraries as I have read in a thread on this forum that the new music scanner might put too much strain on plex with large libraries (hence the cpu load?) with no difference to the result
    All attempts were tried with a backed up database to start fresh. Only sulution to have a working server is to uninstall and install 1.17 again. I might even have tried other approaches and combinations/variations but might have forgot/omitted them. Windows firewall and malware scanner is also disabled the whole time.
    When upgrading and deleting the database plex runs normally, however (of course) all libraries, users, watchstates etc are gone. i have invested quite a lot of time on matching movies and maintaining correct watch-states etc. Starting fresh really is not an option and quite frankly should not be option.
    I have attached my logfiles above and any suggestion on how to tackle this problem would be very much welcome.

Best regards

3

Attempted a repair both before and after upgrade to 1.18.x via https://support.plex.tv/articles/201100678-repair-a-corrupt-database/ but to no avail. Is my problem most likely stemming from database issues or something different?
I previously used Windows 7 in the hyper-v machine but upgraded to win10. Both installations have the same symphtoms. Any help or direction is greatly appreaciated.

4

Just tried with the latest 1.18.6 version but no luck. Anyone have any suggestions?

5

just tried disabling all plugins except for the subzero plugin as described in this thread: Problem upgrade server from 1.17 to 1.18
I am out of options to try. Any help is mich appreaciated.

6

My Problem:
I run 1.17.0.1841 and when 1.18.0.1913 I upgraded the normal
way but after the installation I can start Plex Media Server
but it will not find my data! I uninstall 1.18.0.1913 restart
my computer and install 1.17.0.1841 again and it works perfect
again. I have tried all versions up to 1.18.5.2260 with the same
result I end up uninstalling 1.18 and reinstalling 1.17.0.1841
and everything works again! Has anybody any idea how to fix this???

Don’t ask me how But now I’m running Version 1.18.6.2348.

There were 2 files:
com.plexapp.dlna.db-wal
com.plexapp.dlna.db-shm
in this directory:
C:\Users\JohnJ\AppData\Local\Plex Media Server\Plug-in Support\Databases
that I deleted and after that it worked to upgrade!!!

Comment on this from OttoKerner:

These two files are created during normal operation of PMS.
The fact that they exist, means that
the server was not shut down (yet) in an orderly fashion
the server has crashed
If you delete them, you will lose changes that were made to the
database after it was started last time.

Comment on this from TeknoJunky:

Hello, some plex upgrades include database changes.
If you have a large library db, and/or a slow server,
it may take several minutes for the library database to
rebuild itself (often you will see 100% on a cpu core and
a lot of disk activity). This is could make the server take
a long time to start actually responding… the db upgrade
happens before, and must complete before, the web server
starts responding…
If you are impatient during this process and try to shutdown
plex or the server, you can end up with a corrupt database.
It sounds like from above, that either plex was still trying
to update the dlna database, or it got corrupted at some point.
In the future, if plex web does not respond right away, you might
want to be patient and wait 5-10 minutes before panicking.

My Comment:

After deliting these 2 files:
com.plexapp.dlna.db-wal
com.plexapp.dlna.db-shm

Plex responded directly with version 1.18.6 installed and hast worked
perfectly for the lasy couple of days.

7

Thanks for the input. I have read about this approach in another thread. Unfortunately I have already tried this previously. Disabled DLNA before upgrade and also deleted the dlna Database files. Will try again with dlna Database file kept untouched right now. Have also tried every 1.18.x version previously. Seeing success story’s from others gives me hope :slight_smile:

8

Just took a look at the library.db file. Here are the contents of the sqlite_sequence table:

stream_types	3
accounts	24407871
plugins	72
plugin_prefixes	272
sync_schema_versions	1
tags	160639
play_queues	5071
library_sections	12
section_locations	13
directories	16626
metadata_items	409924
metadata_item_accounts	317469
devices	81
statistics_bandwidth	49668
media_items	122430
media_parts	123388
taggings	2685267
media_streams	232453
metadata_relations	34318
play_queue_generators	483070
play_queue_items	44492
metadata_item_settings	321859
statistics_media	323115
media_part_settings	64
metadata_item_views	326973
preferences	3
media_provider_resources	4
media_subscriptions	8
synced_metadata_items	8178
synced_ancestor_items	897
synced_library_sections	21
media_grabs	1
media_stream_settings	72426

Maybe someone can compare the sequence table to reveal major differences in hopes to dig deeper into what table might be the problem (if indeed related to database problems). Starting 1.18.x with a deleted .db file starts up the server just fine and the settings can be accessed and new libraries can be created.

9

I have tried to upgrade to 1.18.7 which released rescently with no luck.
Logs.zip (18.1 MB)
Can someone please take a look at my Logs and see whats preventing me to upgrade this whole time.

10

I have tried every PMS version which is out up until now. None of them worked.
What I found was that when I removed every Media Library and installed the update and rolled back on failure I coud now figure out that one mediaLibrary in particular was causing this issue for me. Out of the 9 Libraries I created one Movie Library is causing my issue. When I remove it before updating everything works fine.
I have already repaired the database before and after updating and in between.
How can I find which movie, which setting or which corruption in this medialibrary might cause this issue.?
Is there any integrity check I can run?

11

I have made a copy of my virtual machine and removed every library except for the “broken” one. Somehow the upgrade to 1.18.x worked after trying things i cannot recount right now.
Still the size of the .db file stayed the same. Is this normal?

12

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