It would be great if the Transcoder Settings had configuration to allow for the “Optimize” queue to support more than one processing job at a time.
Currently I run Plex in a distributed manner in which actual transcoding jobs get offloaded to other servers in my network, so technically I could be optimizing multiple items in parallel and just have PMS monitor the progress without actually doing any of the work.
However, the Conversion queue is fixed to just one job at a time leaving the servers idle (except for real-time transcoding which does handle all of them in parallel).
Vote for this feature here:
Thanks @darcilicious! Yeah, I did see that thread, and have voted for it too.
I did end up building my own distributed transcoder of sorts a while ago, which works great for real-time transcoding, but doesn’t get fully used for queued optimizations, which is why I opened this feature suggestion.
If I find the time I’ll document how it can be used and post the link to the github on that other thread, in case anyone finds it useful and has a similar cluster/hardware configuration.
I think this post is in double with Distributed Transcoding/Scaling
I don’t think this is the same feature request.
What is requested is to be able to tell plex that it is allowed to optimize 2 or 3 movies at the same time, because now it is fixed to one movie at the time.
I personally run a 128 core system right now. I upgraded storage recently a bit and wanted to optimize all 4k movies to 1080p and 720p. Right now my servers cpu load is somewhere between 5% and 10%. And the optimizing is going on for days.
If I was able to tell plex that it can run 5 jobs at the same time I would still barely hit 50% cpu load.
Better even would be to be able to define certain timeframes of the day where 10 jobs can run while it should pause or run just one or two at other times.
I’m sure by now you’ve completed your queue of optimizations, but this is still in my opinion a must have. I’m going through the process now after losing all of my library and starting over.
I also think that if a movie that doesn’t have an optimization yet is being transcoded live it should save that transcoding as an optimization. This would be awesome for popular movies you know are going to get playtime as soon as you add it to the library.
In 2020 it’s absurd to
a) limit users to using a tiny fraction of their resources
b) assume that using more than one core will always impact on foreground processes
c) not allow users to decide how to use their processing power
I would vote for this but I don’t have a vote…
I would like to see Plex expand this to other process like scanning…
Plex checks CPUs usage and tries to use 80% (user setting)…
My systems all have 4 cores and Plex doesn’t make a dent unless transcoding and since these systems are dedicate to PMS they are greatly under utilized, It also causes unnecessary uptime because it takes so much longer for the process to complete. (my servers sleep when not in use)
Yea, if you’re sitting there pushing the buttons I have to assume you would know if your Plex server is busy or not - and since you actually want an optimized version before the turn of the century, you’d like that to happen as quickly as possible…
But if the minds at Plex haven’t figured that out already - this feature request is a waste of time.
I have a vote left - but Handbrake does all my ‘Optimizing’.
I spend most of my life creating material Plex won’t transcode - the last thing I want is to actually have Plex transcode something.
Juice I do agree with you…
Any thing that you can do yourself and not have Plex do it is a plus. Since most things seem uncompleted, half ass I would say, and seems like no need on their part to even try to make things right.
But there are a lot of things that do need to get done in the background and I would hope that it would get done in a timely fashion. Utilizing 1% of the CPU is not timely…
“Options” are something Plex removes from the ‘user equation’ with maximum speed.
Not all users are as dumb as a bag of hammers - and some might even know when their server is too busy for a user selected job - but we’ll never have such options until Plex discovers we can actually make some decisions on our own without Plex’s help.
The main point is they have gone for a one-size-fits-all approach. But one dated about a decade ago.
I have 32 cores. Why would I buy good CPUs if I can only use 3% of its capacity on CPU-heavy tasks?
If you read the forums you see that the majority of people of Synology or qnap NASes or other low powered gear to run PMS. For them the decision is still the right one.
Still, a switch would do no harm I would say.
The default action could be ‘low and slow’.
Plex could even hide the ‘Dam The Torpedoes - FULL SPEED AHEAD!’ button behind intricate dropdowns (you can’t keep open if you’ve lived more than 20 years) - like Clean Bundles, for instance - so nobody will actually stumble over it.
…BUT NOOOOooooOOOO!
Best to have everything dumbed down into ‘Slime Mold’ territory, so nobody makes any mistakes…
I totally disagree, since I think it would greatly help them to use the idle time to process things. And that would in turn speed up those processes to completion.
I run a number of Synology NAS units that are not totally utilized…
Agree on this. Still you can easily lock up a low powered NAS completely if you run a process utilizing all cpu power at once.
You can lock up a Big Nasty Gaming Rig too - if you know what you’re doing…
Try about 10 remuxes, an Xmedia Recode job, while Handbrake is running on the same drive with the database - and ‘slowdowns’ (read: Hard Aground) are a real and present danger…
![]()
That is true, that is why Plex needs to monitor it’s self and why i suggested a setting like 80% utilization that is user adjustable. Only the admin of that equipment knows what it is used for and should set it accordingly… and nothing over 80% should actually be an option mainly do to unknown CPU spikes from other processes that could lock up the unit.
Does ffmpeg even have a switch for limiting cpu percentage? Plex only calls it got doing the job so it probably has to implemented there.
Edit: Answer: ffmpeg only allows to limit number of threads to use. Fair enough, one could select how many threads one wants to use.
You can actually fry a hard drive…
We used to do it with outdated hardware just to see what it could take…