I am running the latest version of PMS on Ubuntu 18.04 server. I noticed that for most things, even like hour long videos the entire thing is sent to the client (usually a phone or ipad) within first few minutes of playing the video. This is great for stability of the stream (in case there is a connection problem). But it chews through my data capped internet if someone is video surfing (watching just a bit of a bunch of episides), as watching for a few minutes uses bandwidth of the entire episde.
Is there any way to limit the amount that can be buffered? Like lets say to 5 minutes?
Its already limited to 4 Megabit. The problem that it needlessly sends over the entire video in just a few minutes. Rather than sending like a few minutes of stuff at a time.
Does the server software on Ubuntu act differently than the server software on Mac? The reason I ask is that I have never observed the behavior you describe. For example, my sister was streaming one of my movies a few days ago, and despite having lots of bandwidth, it only streamed enough to buffer maybe a minute or two at a time. At least, that’s how it appeared when I checked the activity dashboard.
Not sure, never used mac version. Was server actually fast enough to transcode fast enough for her to load the entire thing? I am running hardware transcoder that is capable of transcoding an hour long show in a few minutes.
Yes, the server is a quad-core i7 Mac mini with hyperthreading. To test its streaming capacity, I have had it serving six simultaneous clients before, three of them transcoding at various bitrates and three direct streams. When my sister was streaming the movie I mentioned in my previous post, she was the only client, the CPU was never over 10%, and most of the time under 5%.
The only reason I thought of it when I saw your question here is that I had been wondering when she was streaming why my server didn’t just send her the whole thing at once; it could clearly have done so, but instead it seemed to be ‘pulsing’ it out to her 30 seconds or a minute at a time. The bandwidth chart looked sort of like an EKG of somebody’s heartbeat… low most of the time, with a little spike every 10-15 seconds when it sent out a chunk of data.
I don’t mean to sidetrack your question… but maybe if we could figure out why mine behaves differently than yours, we could get yours to behave the same way.
For what it’s worth, my sister’s stream was also a direct stream. But I have seen the same behavior with transcodes.
One of these clients - PMP/PfW/Roku? - will allow setting the buffer size. I suppose, if conditions and media are just right - an entire file could be put in one buffer.
Like a Looney Tune or Anime Tune - or some super-duper-bit-starved-stuff (even worse than mine).
Typically, by design, Plex pulses along, jacked to 80% of limit until the buffer is filled - then a rest - then go again. Like NASCAR - go like hell, then turn left.
If you’re trying to save your data cap you’re gonna have to limit the remote stream bandwidth globally (server/settings/remote access) to some ridiculously small number and hope that works as advertised. <— and also hope you can transcode crappy streams to as many clients as you’re serving.
