How the hell is “I don’t know” the accepted answer for this?
I find it difficult to believe that this is a caching issue, because even if a sync were performed and the file is actually on the device, the cache IS still used to hold all, or parts of, the file during playback.
Consider the following events in my logs
Apr 16, 2018 09:17:49.946  DEBUG - Completed: [A.B.C.D:port] 200 GET <<== denotes the approximate end of playback for the prev-track
Apr 16, 2018 09:17:56.759  DEBUG - Failed to stream media, client probably disconnected after 131072 bytes: 10054 - An existing connection was forcibly closed by the remote host <<=== notice the ~7 seconds of time that has passed before an error is reported saying that an EXISTING client connection that was expected to be there, is now gone. This implies that the client side of the connection was closed prematurely and it took a few seconds for everyone to figure out what was going on.
Apr 16, 2018 09:17:59.962  DEBUG - Client [2c6dbd1170e3a052-com-plexapp-android] reporting timeline state playing <<=== another ~3 seconds later playback finally resumes after another round of authentication and a new connection can be established.
I was not experiencing this issue previously when I had:
- Win10 PMS install
- apache2 proxy used to handle domain specific web request forwarding
I am currently experiencing this issue after:
- upgrading to nginx for proxy forwarding (still have Win10 PMS install)
- enabling SSL in nginx and via plex
I have a strong suspicion that this issue relates more to SSL connection handling than it does anything else. I will be trying to tweak various nginx settings in the upcoming day or two – if I can identify the issue and resolve it, I will report back here. If any Plex staff wants access to relevant log information let me know.
PS: tweaking nginx settings because no authentication is required between the nginx host and the PMS host via plex settings.