Correction to my previous posting.
The directory structure for /mnt/music is incorrect, it is now /mnt/audio/music.
Ok. that’s better.
The systemd message is being printed because you changed the override.
sudo systemctl daemon-reload to get the date-time stamps updated (internal housekeeping) to make the change ‘official’.
Right, but I’ve already done sudo systemctl daemon-reload multiple times. What I’m saying is that that is being requested of me after every reboot of the system and the change isn’t being made ‘official’ for some reason.
Please run ‘sudo journalctl -xe | less’ and skim through it.
Something isnt’ right with systemd
I wasn’t able to find anything that seemed to indicate an issue with systemd.
However, after I relocated my music library location, PMS had to re-scan the entire music library. At the end of its scan it crashed. Upon reboot, I removed the old Music library location, which initiated another library scan and after it completed it crashed, again.
I turned off every Scheduled Task, and the server did not crash overnight.
Today, I’ve been going through my entire music library’s directory/file structure and cleaning up and reorganizing at the directory level. I spent hours on “A” alone…
I manually ran another scan, once I completed “A.” So far, it hasn’t crashed after the scan.
However, a grep of syslog does reveal the following things:
Aug 25 16:41:33 Plex-on-Pi Plex Media Server[601]: ERROR: Error decoding audio frame (Invalid data found when processing input)
Aug 25 16:41:54 Plex-on-Pi Plex Media Server[601]: ERROR: Error decoding audio frame (Invalid data found when processing input)
Aug 25 16:41:57 Plex-on-Pi Plex Media Server[601]: __code__:181: FutureWarning: The behavior of this method will change in future versions. Use specific 'len(elem)' or 'elem is not None' test instead.
Aug 25 16:55:11 Plex-on-Pi Plex Media Server[601]: ERROR: Empty fingerprint
Aug 25 16:55:12 Plex-on-Pi Plex Media Server[601]: ERROR: Empty fingerprint
Aug 25 16:55:16 Plex-on-Pi Plex Media Server[601]: ERROR: Empty fingerprint
Aug 25 16:55:17 Plex-on-Pi Plex Media Server[601]: ERROR: Empty fingerprint
Aug 25 16:55:52 Plex-on-Pi Plex Media Server[601]: ERROR: Empty fingerprint
Aug 25 16:56:05 Plex-on-Pi Plex Media Server[601]: ERROR: Empty fingerprint
What do these errors imply?
I spoke too soon. PMS has crashed. It just took a few minutes after the scan completed.
Are there any other operations besides those for dealing with a corrupted database ( https://support.plex.tv/articles/201100678-repair-a-corrupt-database/ ) that one can perform on their system?
Things like delete caches, metadata image directories, etc.?
I truly believe my issue is with my database/library. Besides the fix for a corrupt database, which didn’t fix it, what else can I do here?
I need the logs please before any more info is lost
Thank you.
Again we see the hardware / kernel fault.
Aug 25, 2020 17:03:24.589 [0x70dfe440] DEBUG - Completed: [127.0.0.1:33724] 200 GET /library/metadata/33181/tree (7 live) GZIP 27ms 2682 bytes
Aug 25, 2020 17:03:24.642 [0x609fa440] DEBUG - Request: [127.0.0.1:33726 (Loopback)] GET /library/metadata/33083/tree (7 live) GZIP Signed-in Token (derMechaniker)
Aug 25, 2020 17:03:24.671 [0x715ff440] DEBUG - Completed: [127.0.0.1:33726] 200 GET /library/metadata/33083/tree (7 live) GZIP 29ms 2323 bytes
\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00
… 00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00
This binary works for many many others. I’m sorry but there must be a system problem beyond what I can see here.
CPU overheat, SD card, power, cabling, something very very simple.
Help me understand this binary message. Forgive my lack of expertise in reading that type of message.
Is that an indication of a kernel panic?
What, precisely, is PMS working on when that occurs?
While I am understanding the idea of a possible issue with CPU overheat, SD card, power, cabling, etc., I am having a hard time understanding how that could be the issue.
The reason why I have a hard time accepting that is because the way in which the system crashes seems to be related to a scan of the Music Library itself.
We see this crash happening at the exact same time during AM Scheduled Tasks. If this were an overheat, power, cable, SD card thing, it would have a more random-like failure. It would be occurring during other activities on the server. It would extremely hard to have a verified repeatable crash occurring within a 1 minute tolerance on a 24 hour cycle.
I can force the crash by just performing a manual scan. It’s that easy to reproduce.
Furthermore, I’ve replaced the Pi, the SD card, re-installed the OS & reinstalled PMS.
I’m using a high quality power supply, which I can swap with another identical unit.
I cannot easily replace my 4-bay SATA enclosure that my SSDs are in.
With all my years of hardware/software troubleshooting, I really believe that my database/library’s integrity is the issue here.
Frankly, I’m at the stage of accepting that my database is probably going to have to be totally deleted and started over fresh.
As I said in an earlier post, I was going to try generating a new Plex database, add a bit of content to it and let it run as normal. If it does not crash, then clearly the issue is my data.
As expected, a new PMS “test” database with minimal imported music content ran without issue.
Yesterday, I decided to completely start over from scratch, accepting that I would lose all the hand-touched metadata editing I’d done to my database within PMS over the years — technology can really suck, at times.
- Format SD card
- Install fresh Raspberry Pi OS Lite
- Install fresh Plex Media Server
- Configure fstab
- Configure override.conf for PMS
- Import Music into PMS
- Wait for early AM Scheduled Tasks and observe
At this stage, the server has not crashed, and has successfully completed its Scheduled Tasks.
I’m cautiously optimistic.
I have backups of my old database dating back to 2018, but I’m reluctant to get into seeing how far back a stable backup may be, nor do I want to upset the balance I have finally found after, at least, a month of this, near daily, nonsense.
The most disturbing part is not understanding what happened to cause this to have occurred in the first place.
This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.
