Seen this a few times. Latest beta.
Shows two streams from same IP, same user. (This guy is sitting two spots down from me and I verified he’s only watching the one)
Plex Media Server Logs_2020-01-16_12-16-06.zip (5.7 MB)
Did he possible open up a second Plex Web window after starting playback? This will bring up the same video paused in the mini player.
Your screenshot shows the second one (buffering) being 2 minutes behind.
Just asked.
He started an episode, minimized it, went to a different episode and played that one. All within same browser window.
Sounds like the first stream failed to terminate.
I do see on the screenshot it’s the same episode but I think that’s more indicative of the issue if anything, since they should be different.
I’m not able to reproduce. Did you have the dashboard open the entire time? Wondering if Web just needed to be refreshed?
Nope, was watching something myself and just went to check.
Yeah, it’s fairly rare, but I’ve seen it a few times now.
Only difference is the user isn’t across town. I can yell at him to troubleshoot lol
@anon18523487 FYI, just happened again. This user never clicked off of that window.
Plex Media Server Logs_2020-01-17_11-22-32.zip (5.0 MB)
These are also about 2 minutes apart. I wonder if something happened while streaming to cause a glitch so PMS started a new stream, but the old one didn’t get cleaned up properly. I didn’t see anything in your previous log, but I’ll check this new one.
Not sure?
The two users are here at work (200gb egress from our network) and the server at home is fiber as well. Latency is 3ms or less usually. (Major research university)
IO, CPU, Mem on server are way overbuilt.
Trying to think of anything from bare metal up that could cause it but no clue?
Clients connect to the server using Websockets. These can sometimes fail on their own. Plex will establish a new connection and keep going, so as a user, you wouldn’t notice. However, on the back end, the stream does change.
Also, web isn’t very good at cleaning up after itself in these situations. A dropped connection like this could appear the same as paused playback. I’m not sure there is a fix. In your PMS settings, there is an option to stop playback after being paused for a settable amount of time. Enabling that option might clear these from the dashboard after the time you set.
Sure. I have that set to 30 minutes IIRC.
It’s not causing a strain or anything, so I’m not sweating it too hard.
Just was odd enough to kick up to you folks.
Thanks for the report. If you notice it does cause an issue please report that and we can take a further look. For now, I believe this is just aesthetic and shouldn’t cause any problems. Not sure it’s worth the devs time right now to track this down.
fwiw I have seen the same thing before but unable to reproduce it on demand.
I’m cool with that.
If a lower powered server / network combo was hit with it… maybe? If that’s the case though, they’d be used to issues.
I’ll wrap some logging around it via cockpit and ping you if it creates any noticeable resource issues. From what I can tell though, it just sits idle as an open socket. No bandwidth or cpu.
I believe that is what is going on. Websockets do take up a small amount of memory but it shouldn’t be noticeable.
I’ve been able to reliably reproduce it by starting playback via a web client and then refreshing the web page. It will show a duplicate stream, which eventually ends on its own after the “real” stream is completed. As far as I can tell, the duplicate is not using much in the way of resources, either CPU/memory or network (maybe some negligible amount).
Ah yeah, that would make sense. When you refresh Plex Web, it basically drops the current connection and starts a new one. This doesn’t send a close signal to PMS so the websocket stays open.
1 Like
This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.
