Arch Linux

justdanyul

Member

Registered: 2011-09-29

Posts: 130

Plex media server upgrade steps

Hi all,

I recently tried upgrading my plex media server for the first time, I basically just upgraded it like any other AUR package, but oh my got it got angry with me...

Long story short, after the upgrade, the systemd service refused to start again. I ended up having to remove plex, and manually delete everything /var/lib/plex, re-install and basically start over with a fresh install as per the wiki page.

I'm just wondering, how come it exploded so bad? what is the normal process for upgrading plex media server?

kcrmson

Member

From: 20 min from the Cobra Kai dojo

Registered: 2018-06-01

Posts: 12

Re: Plex media server upgrade steps

That, in my opinion, is pretty unusual, possibly due to maybe some sort of database corruption or even incorrect permissions (user:plex group:plex, 755 perms on most of the stuff in /var/lib/plex and your media directory/directories.

I'm using the plex-media-server-plexpass AUR package myself (shouldn't differ from yours in the case of it just starting the service), my routine for updates is pretty streamlined (this is while the server is still running as well, I have Plex Media Player open either on the host machine or another machine to keep an eye on it while updating):

Update repos and packages (-Syu)
Either use an AUR helper (yay in my case) or build/install the package yourself (makepkg -si)
sudo systemctl daemon-reload
sudo systemctl restart plexmediaserver.service
Give it a few seconds for the previously connected Plex Media Player to re-establish its connection, it might even say insecure at first but if your NAT and port forwarding is setup correctly (e.g., not double NAT) it should quickly switch to being secure and your home page should show up as expected.

I've been running mine for the last few years like this, nary a failure on an update. I'm still thinking your issue (since you started it fresh getting current logs isn't going to help) might be explorable if perhaps you can pull up anything from journalctl from around the time it failed to run again. Suspecting permissions or weird database corruption myself, though corruption on a fresh server would be extra oddball unless you had some underlying hardware issue or a loose SATA cable or something of the nature.

-kc