Server Version#:1.28
Player Version#:
We really need more info here to be able to help, as well as PMS Logs in debug mode
Most of these āupgrade killed the serverā posts are database corruption.
I appreciate that and have looked at the logs indicating some corruption. Some Db files have also changed in size significantly (cf to a copy working server) and will attempt to remedy (not optimistic). My greater concern is that i (foolishly) updated 3 different machines with this update and all have the same problem. From working Plex servers to ones that dont work. The process appears to damage or remove large parts of the DB files.
What version did you upgrade from?
This is an incorrect statement, prior to the upgrade the databases where already damaged. During the upgrade process Plex ran into the corruption, and tried to deal with it. GIGO, Garbage in, Garbage out.
Even though you are correct in what you say, thereās also the option of upgrading from a very old version, where converting can take a very long time, and was interrupted by the OP, if loosing patience 
Thus why I asked for what version it was before
True 
The history of upgrading:
Synology 3617xs on DSM 6 upgrading from 2 revisions prior to 1.28 . Upgrade failed with common symptoms and corruption error.
I took the opportunity to upgrade the unit to DSM 7 reinstalled plex from a recent backup and reinstalled (upgraded) plex using 1.28.0.5999 - failed with corruption error.
Removed 1.28 and entire plex folder and reinstalled using previous back up.
reinstalled plex using 1.26.1.5772 - System works fine.
Upgrade a Synology DS918 running DSM7 with verison 1.26.1.577.(working nicely) to 1.28 - Failed with same symptoms and corruption error.
Upgraded docker on unraid server yesterday (working nicely) to latest version failed with same symptoms and corruption error.
3 different machines.
Not disgareeing there may have been some errors in the originating database (all3?) but suggest plex would not be working if they were there on the earlier versions.
Most database corruption is caught during the upgrade process as that is the point where plex has to touch the entire database (table upgrades, adding/removing indexes, ect) Most of the time plex can operate with a corrupt database, some features may not work, or some records may be missing/unable to be updated, ect.
I will note that unraid has a significant rate of database corruption.
An easy way to tell if the database has issues prior to upgrading is to do a PRAGMA integrity_check and see if it reports issues with the original working database/plex version.
Help! I currently have PlexMediaServer-1.27.0.5897 installed and when I attempted to update Plex Media Server to 1.28.0.5999 (via App Central), the install hung just at the end of the installation. I let it go for about an hour and it appears to have timed out, and then āupdateā option appears again. When I attempted to update again, it just appears go through the update, then comes right back with no errors - still showing 5897. I could not open Plex Media Server following this update attempt. I tried restarting server and reinstalling with no success.
I also tried manually installing 5999 and 5897 with no success - same result.
I also tried disabling the app under āInstalledā apps in App Central, but now I get the message when I attempt to re-enable: āThe action cannot be completed because there is already a plexmediaserver folder inside the Web shared folder. Please first remove or rename the folder before trying this action againā.
Are there any suggestions for ārefreshing/resettingā the current installation without having to remove the media server? I am hosting on a Asustor AS5202T.
Here are the first logged errors when the initial install attempt failed; all subsequent attempts fail with same error:
08/03/2022 AM 10:08:34 ERROR admin plexmediaserver Plex Media Server-1.28.0.5999-97678ded3 Install App failed. (Ref. -2)
08/03/2022 AM 08:20:38 ERROR System plexmediaserver Plex Media Server-1.28.0.5999-97678ded3 Unknown error (session timeout)
Previous entries were successful going back to 2019:
07/30/2022 PM 01:47:47 INFO admin httpd-2.4.43 Apache HTTP Server-2.4.54.r16 Install App success.
These are the message/errors logged when I tried to disable and re-enable:
08/03/2022 PM 06:20:40 ERROR admin plexmediaserver Plex Media Server-1.27.0.5897-3940636f2 Enable App failed (unknown). (Ref. -161)
08/03/2022 PM 06:20:34 INFO admin plexmediaserver Plex Media Server-1.27.0.5897-3940636f2 Disable App success.
I had the same exact thing happen when updating to 1.28.0599 on my Asustor.
This version seems to be corrupting the database.
I got stuck on 5897 and it wouldnāt update either. I also couldnāt open Plex but after an uninstall and reinstall it came back so I just left it on 5897. Maybe I got away without corrupting my database but it sounds like you might not have gotten so lucky.
Youāre probably going to have to repair the database but you might want to wait for an employee to verify that.
If youāre feeling brave you can follow the instructions ChuckPa gave in another thread.
On another note, iām curious now
Backing up the database from an older version now, does this mean a corrupted database will happen again in the near future from another update?
The biggest problem I run into here in the forum is āpanicā.
One very common scenario is:
- Install an update before reading the release notes
- Expect the server to immediately be available again
- Become impatient and FORCE STOP the server
- Restart again
- When it fails again, they go digging
- Itās now discovered the DB is corrupt.
How did this happen?
The most frequent scenario, which Iāve caught myself ALMOST doing, is not checking to see if the CPU & the Plex Media Server process are
- Running
- IDLE (near idle)
They donāt check, NOT KNOWING there is a database update in progress.
Some DB updates are so big they take a few minutes even on big machines.
(It all depends on how much media is indexed by PMS)
The result is they forcibly terminate PMS ā MID UPDATE ā and leave the DB file in a completely corrupted state
Whoās to blame here? Impatience.
How to prevent?
- Read the release notes
- Check to see if PMS & CPU are really idle before acting.
NAS boxes, with their smaller CPUs are VERY prone to this. I see it on Synology all the time.
I have a dedicated FAQ for it
Thatās all.
steps off soapbox
1 Like
Updates do not cause database corruption on their own, has ChuckPa says not giving the database enough time post upgrade, or things like loosing power unexpectedly are what normally end up causing the corruption.
Anytime you suspect you might have corruption or if you just want to paranoid, you can do a PRAGMA integrity_check before any upgrade.
Appreciate your observations. I have been a Plex user for over 10yrs now so am pretty attuned to the nuances of the program, its limitations and its updates. Settling in to work through this problem like the many before.
Thanks for you help! I was a little paranoid of uninstalling the app and re-installing for fear of losing metadata, as I have not done this before - but it worked, and I was able install 5999 from the App manager! Looks like everything was intact following install!! I do not think I had any corruption checking disk health and logs, but I did notice following the first install attempt, the disk usage was constant after I restarted - so I know it was doing a ton of updates.
This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.
