Cleaning up the plugin directory

I have a PMS installation with a large config directory (60GB+) that I have copied from machine to machine over the years of running Plex. It has a huge amount of effort involved into the metadata, so I don’t want to lose it.

Recently I have started to experience occasional PMS hangs for no apparently reason. I have tweaked settings and looked through logs with no real luck to pinning down the cause. I posted logs on forums with no luck there either. No one else seems to be having these types of hangs, so I figure it is something unique to my setup.

I am migrating from Plex running under Debian to the Docker version of Plex, all on a high end dedicated server. My hope is by moving to Docker these hangs will go away, but I have a fear they’ll continue because my installation was pretty standard, and of course I am still using that same configuration directory.

I noticed I had some very old plugin versions (things like Google Music, Unsupported App Store, Pandora, etc) in my plugin directory that I never use. In my desire to clean everything up, I removed all these plugins via the UI. However, my plexmediaserver/Library/Application Support/Plex Media Server/Plug-ins directory still shows this:
DevTools.bundle/ Services.bundle/ SiteConfigurations.bundle/ WebManager.bundle/ master.zip*
The first 4 are directories, the last one is a file. Should I delete these? Leave them as-is? Replace them? They have dates from 2014 (except for Services.bundle).

Any other suggestions on “cleaning up” my installation? Other than blowing away all the metadata and starting from scratch, of course!

You need Services.bundle. The rest are optional / your choice but not needed.

With regards to cleaning up the metadata, have you reclaimed any space by doing the ‘Empty Trash’ and ‘Clean Bundles’ steps?

Yep, thank you for reminding me of those steps. Unfortunately they didn’t cure my problem, although I eventually did. My PMS metadata folder was on an NFS share to a fileserver. I had enabled file locking on the share and had verified it was working. However, once I moved the metadata folder to a local drive (had to install a new one to do it), it eliminated the hangs.

Yes, local_locking does need to be posix in 99% of the cases. This is the only thing I’ve found to reliably allow the DB to be on the NAS and PMS on another host.