Server Version#: 1.16.3.1402-22929c8a2 and 1.16.3.1433-359b06978
Player Version#: Version 3.104.2
Yesterday I couldn’t scan library files. I found these errors just after starting plexmediaserver, it happens withing 2 minutes of service start.
Jul 25, 2019 12:29:47.876 [0x7f0b367fc700] ERROR - SQLITE3:(nil), 1, statement aborts at 1: [COMMIT] cannot commit - no transaction is active
Jul 25, 2019 12:29:47.876 [0x7f0b367fc700] ERROR - Failed to commit transaction (…/Library/MetadataItem.cpp:958) (tries=9): Cannot commit transaction. cannot commit - no transaction is active
Jul 25, 2019 12:29:49.002 [0x7f0b367fc700] ERROR - SQLITE3:(nil), 1, statement aborts at 1: [COMMIT] cannot commit - no transaction is active
Jul 25, 2019 12:29:49.003 [0x7f0b367fc700] ERROR - Failed to commit transaction (…/Library/MetadataItem.cpp:958) (tries=10): Cannot commit transaction. cannot commit - no transaction is active
Jul 25, 2019 12:29:49.003 [0x7f0b367fc700] ERROR - Exception inside transaction (inside=1) (…/Library/MetadataItem.cpp:958): Cannot commit transaction. cannot commit - no transaction is active
Jul 25, 2019 12:29:49.003 [0x7f0b367fc700] ERROR - SQLITE3:(nil), 1, statement aborts at 1: [ROLLBACK] cannot rollback - no transaction is active
Jul 25, 2019 12:29:49.004 [0x7f0b367fc700] ERROR - Failed to roll back transaction (…/Library/MetadataItem.cpp:956): Cannot rollback transaction. cannot rollback - no transaction is active
Jul 25, 2019 12:29:49.004 [0x7f0b367fc700] ERROR - * Error loading episode.
Jul 25, 2019 12:29:49.036 [0x7f0b367fc700] ERROR - Failed to commit transaction (…/Library/MetadataItem.cpp:958) (tries=1): Cannot commit transaction. out of memory
Once this starts in the error logs, the media scanner will never finish or find new files. If I stop the server and restart it, it will find a file or two, then go back to this. Playback and other server functions aren’t apparently impacted, just this.
I am running on Ubuntu 18.04.2 LTS, Linux 4.15.0-55-generic, 12 CPU KVM VM with 16Gig of ram.
The “out of memory” seems especially odd, as I’m well under 50% system RAM utilization.
total used free shared buff/cache available
Mem: 15849 1276 6999 7 7573 14235
Swap: 30515 0 30515
I was running 1.16.3.1402-22929c8a2 when this started. I followed the “Repair a Corrupt Database” instructions, sqlite3 gave me “Error: database disk image is malformed”, so I did a dump and import, but aside from the integrity check coming back clean, it’s the exact same problem.
I upgraded to 1.16.3.1433-359b06978 just now, hoping it might have had a bug fix, but it’s exactly the same.