Server Version#: v1.19.3.2843-e3c1f7bcd
Player Version#: n/a (server crashes repeatedly, not available to clients)
This is the 64-bit linux version of PMS installed on an Ubuntu 16.04 VM, which has been allocated 4 cores, 4GB of RAM and a 40-odd GB virtual disk, of which about 35% is used. There’s nothing special like hardware transcoding involved - this is all vanilla CPU-powered.
It’s run fine for a long time (hence 16.04 rather than the two newer LTSes), up until yesterday when I updated to the new server. Since then, all it does is repeatedly segfault when started.
syslog shows repeated occurrences of:
May 22 09:31:12 media Plex Media Server[1820]: ****** PLEX MEDIA SERVER CRASHED, CRASH REPORT WRITTEN: /var/lib/plexmediaserver/Library/Application Support/Plex Media Server/Crash Reports/1.19.3.2843-e3c1f7bcd/PLEX MEDIA SERVER/4368db0b-2e07-1f42-14681723-2cdfbaed.dmp
Since the web UI is unavailable, I can’t go that route for providing logs. If needed, I can produce a tarball of the Logs directory. I’m assuming that crashdumps are already available as the uploader has been working overtime.
At this point, the missus and the kids are grumbling at me as they can’t get at their music and the likes. I’m probably going to spin up a new blank image and install from scratch, but I can keep this instance around in case more info is needed.
How do you want those delivered - just as a tarball of the complete contents of /var/lib/plexmediaserver/Library/Application Support/Plex Media Server/Logs ?
Proceed to setup a new server – However – give it a different “Friendly Name” so we can distinguish it later if needed.
If there is a problem with the host, this installation will also fail.
If there was a very deep problem in the DB (which might be given the database recovery statements in your logs) this will correct it.
May 22, 2020 10:11:43.966 [0x7f5bf1525740] DEBUG - FeatureManager: Using cached data for features list
May 22, 2020 10:11:43.967 [0x7f5bf1525740] DEBUG - Opening 20 database sessions to library (com.plexapp.plugins.library), SQLite 3.26.0, threadsafe=1
May 22, 2020 10:11:43.968 [0x7f5bf1525740] INFO - SQLITE3:(nil), 283, recovered 333 frames from WAL file /var/lib/plexmediaserver/Library/Application Support/Plex Media Server/Plug-in Support/Databases/com.plexapp.plugins.library.db-wal
May 22, 2020 10:11:44.005 [0x7f5bf1525740] DEBUG - Running migrations. (EPG 0)
May 22, 2020 10:11:44.123 [0x7f5bf1525740] DEBUG - ChangestampAllocator: initialized to 21567475
May 22, 2020 10:11:44.123 [0x7f5bf1525740] DEBUG - Opening 2 database sessions to library (com.plexapp.plugins.library.blobs), SQLite 3.26.0, threadsafe=1
May 22, 2020 10:11:44.125 [0x7f5bf1525740] INFO - SQLITE3:(nil), 283, recovered 537 frames from WAL file /var/lib/plexmediaserver/Library/Application Support/Plex Media Server/Plug-in Support/Databases/com.plexapp.plugins.library.blobs.db-wal
Yeah, it looked pretty impressively stuffed from what I could tell. However, I’m not familiar with the software from anything but a user’s perspective, hence this thread.
So, to be clear, by moving Library/ out of the way, this is effectively starting a new server from scratch, right? This isn’t a big deal in the grand scheme of things; I’d planned on just spinning up another VM entirely from a blank image I have lying around.
Either way, it’s just that the rescan time will be a pain in the bum. However, if it’s conclusively borked, then so be it.
I’ll admit that I don’t really have a backup - my last snapshot of the VM is too old to be of a huge amount of use. I’ve been more worried about the data than preserving the VM itself, and the last snapshot I have dates from late 2017 when I dragged the system up from Ubuntu 14.04 to 16.04.
I really should look into sanoid or similar to handle automatic zfs snapshot management…
Right; Renaming Library to Library.old shoves it off to the side. This is because PMS is looking for Library. If it doesn’t find it, it creates the directory structure. Once it sees it needs to create structure, it knows it is now in a “startup” task and proceeds from there.
rather than all that expense and headache, of ZFS snapshots / VM backups / whatever, try to keep it simple.
Once you get it rebuilt , routinely stopping Plex and making a tar ball of ./Library gives you both a mirror snapshot and a completely portable image in one shot.
You can take that file anywhere and untar it to recover the exact server .