QNAP TS-453e, not run in a docker or with any other add-ons.
When invoking the Metadata Refresh for the music library, the server begins the process and eventually crashes. Watching the alerts screen, it gets through the scanning process and starts the metadata downloading/updating process. The alerts then say that the connection to the server has been closed, and I’m unable to access the server until I restart it.
I’ve done all of the database checks I can find: optimize, clean bundles, I’ve ran the github Plex DB Repair (no issues found with the databases, fully completed the repair/optimize/re-index steps). I removed the last files that Plex claimed it was working on when it crashed from the server, no change. I’ve restarted both Plex as well as rebooted the NAS several times. I’ve looked at the crash logs, but can’t discern anything useful from them. Most likely because I’m not knowledgeable enough to know what to look for.
The other libraries can do a full metadata refresh without any issue. I’m not entirely sure when this began, within the last month or two most likely. I don’t believe there were any other changes made to either Plex or the NAS, just adding files or changing file metadata through another program (mediamonkey) like I’ve been doing for years.
The library is fairly large, 128438 files, 3.5TB, so I don’t know if that’s playing a part in this. I’ve seen others with larger libraries though.
Other than this library-wide metadata refresh failure, Plex seems to run without any other issue. Refreshing metadata at the artist or album level seems to work fine.
If you go have debug logging enabled on the server, then after you stop/start PMS in the App Center and recreate the problem, you should be able to go into a Plex GUI app into the Settings → Troubleshooting section and find a Download Logs button. Clicking on that drops the log files onto your local computer where you should be able to find Plex Media Server.log which you can open in a text editor and search for the word ERROR and possibly just look near the bottom of the file for strangeness. Zipping and uploading the log file here is a very useful step for solving issues, but I understand if you are reticent. A crash like your describing could a be database issue that DBRepair can’t recover, or it could be a bad file on your NAS. In my case once I had a memory stick starting to fail, causing flipped bits at rare intervals, something I only realized after booting to a UEFI USB stick and running Memtest86+. Let’s hope the log file captures something.
Like you said the crash logs aren’t helpful, but the server log generally is, and after restarting PMS, the logs get rotated leaving a very clean Plex Media Server.log that shows how well PMS started up, how well it completes the problem refresh, and any errors along the way.
I manually stopped and restarted PMS. Reran the manual metadata refresh, and it crashed. In order to get to the Download Logs section, I had to manually stop and restart PMS from the NAS app center. I don’t know if that will clean the logs out of what caused the crash.
Prior to uploading the .zip, is there any scrubbing I need to do of personal information from the logs?
Good question. I usually open the log in a text editor and search for things like my email, username, computer host name, the names of any .mkv it might have processed. I replace those with XXXXXXX.
Most of my bugs reports involve rebooting and testing only one video file, leaving me with very little to redact. In your case of a Refresh Metadata issue, it could be a lot of filenames.
I also search the log file for keywords like ERROR, CRITICAL, malformed to give me an idea of what might be wrong. If you see a database problem from SQLITE, we’d know what’s wrong without you having to upload.
I appreciate you giving me some things to look out for in the log. I went through it for those keywords, and found a handful with errors. Some were simple things like an album type not recognized, which weren’t terribly important to the issue at hand but were still nice to fix. There were others that while I didn’t understand what the errors were, I replaced the files with new ones anyway. I’ve now ran two full metadata refreshes without issue.
Thanks for the guidance. Seems to have fixed the issue, and if it creeps up again, I have more of a starting point to try to fix it before coming here.
