In the past during the maintenance window I would hear much disk activity as I assume it performed some quick checks and then the disks would go back to sleep.
About a month ago I redid my whole library and split up the media into separate drive letters so I have new libraries per say but with all the same data.
What I notice is that now during maintenance window the disks are run constantly for the entire time of maintenance, so it appears that it is doing something that it can’t finish in the 3 hours of default maintenance time. I have tried setting the maintenance to 6 hours instead and now the disks run constant for 6 hours. Will this eventually quiet down? it been over a month now since redoing the libraries… (70TB of data)
I think that depends on how powerful the system is you’re running PMS on and what PMS is allowed to do during scheduled maintenance. If you have Generate video preview thumbnails turned on then it might take a lot of time (and a lot of disk space) to generate these for all your video’s. If it is doing anything like that you should be able to see that at the Heartbeat icon in the top right corner of Plex Web.
It will be done eventually, but since you have 70TB of data I assume that it might take a long time. You can maybe check the total size of the Application Support folder (which contains all the meta data and databases for Plex). If that folder is increasing in size during maintenance then I’m pretty sure it’s still generating thumbnails.
I do have preview thumbnails and credit detection turned off for each library. I guess i’ll just set the maintenance window for 12 hours and see if it eventually calms down…
Here’s a link to my current setting for maintenance…
“Perform extensive analysis during maintenance” will have to read all of your 70tb of media in it’s entirery. Looking at your settings, that’s probably what’s taking time (this won’t show up as an activity in Plex Web).
If you measure your read speed on your disks, you can calculate how long time this will take.
Plex is not very optimized here, and it can cause severe read activity on the disks, because all of its activities is ran separate from each other.
Perform extensive analysis: full read of all files
Thumbnail creation: reads all files again (not full file reads, but almost)
Chapter thumbnails: reads all files again (small read amount, but it still needs to go through them again)
Credit detection: reads the tail end of all files again.
Intro detection: reads the first half of all episodes again.
Etc.
Combined, on a 70tb library, the read activity is enormous and will definitely cause wear on the disks. I know this myself, and I very much would wish Plex could combine some of its activities both to reduce the sheer read amount but also to save time.
What is the purpose of perform extensive analysis? Why does it need to read all the data? And I would assume once it does it would stop? Will it run again at some point?
That specific feature analyses how much bandwidth is required to play each file.
If turned off, files that hasn’t been analyzed will use it’s reported bitrate x2. This is fine but can be inaccurate in some cases. The only way for Plex to know the exact bandwidth requirement is to read through the entire file.
So, you can safely turn it off unless you intend to use your server in more bandwidth restricted scenarios, and even then it’s only to make it more accurate.
It will not analyze the files again unless they change, but I think (though I could be wrong) that it will do it again if you move them.
