[Brainstorming] Scan media library task

[BloodyIron 24]

Hi Folks,

 

I just did a search before posting here, as I wasn't sure if this had been discussed recently, and I didn't find anything.

 

I was hoping we could revisit streamlining of the task "Scan media library".

 

My biggest gripe with scanning my library once I drop something in it is how "long" it takes. Now, you might say to yourself, set it up to auto-scan when something is put in, but Emby is using a folder that has an SMB mount, so I don't think it can pick up on new media being present.

 

Now, I'm not an Emby dev, so I don't know the exact way it operates currently, as such I am basing these thoughts on my observations.

 

What I see is when I start a scan, I don't see any of the "new media" until the scan finishes completely. What I hope we could do is have the task broken down into "stages" behind the scenes. First stage just scanning for new media, getting initial names, and requesting the media that a user would experience first. Think about it from a User Experience regard, let's get info that the user is likely to experience first. Name, cover image, something like that, then work through the other things such as Actors, chapters, etc. Also, if a new "season" is detected, grab info for the episodes in order.

 

Now, I suspect one thing that's going to be said is that this process relies on external servers for content. Perhaps that is the slowest part, and perhaps we may not be able to streamline it more than it already is. So be it. But, if we could, at some point (not high priority) revisit the "order of operations" for the scanning, I am optimistic that _perhaps_ we could result in some improvements here.

 

Either way, I'm hoping we can talk about this. Thoughts?

[Koleckai Silvestri 1150]

I feel the proper solution would be to get Realtime Monitoring working over SMB. Even if that is a script or utility that communicates with Emby via the API and is triggered by Crontab. The library scan shouldn't really be run just to add new items to the system.

 

That said, look at your plugins and whether you're extracting chapter images (per library setting) during the library scan. Reducing these loads can make the scan go faster.

[Luke 37060]

Yes, reducing enabled metadata features will make the library scan run faster. That is the trade-off that only you can decide what's best for your setup.

[dcook 265]

I suggest having the scan library task setup per library.  

 

I have 5 libraries and only 1 of them changes every day, the others change maybe a couple times a month if that.  

 

There is no reason to scan everything all the time, when it can be broken down per library.

Edited March 23, 2017 by dcook

[BloodyIron 24]

Ahh I'll try turning off extracting chapter images during library scan, see how that goes Thanks!

 

But yeah, scan per library would be nice.