Server Version#: 1.18.2.2029
Player Version#: Roku player- what ever version is most recent- automatic updates
When playing a TV Episode accessed from on deck the Episode stops playing at 2 minutes until the end of Episode. Has been this way for few Plex versions. I thought at first it was related to auto play of next function but I turned that off and issue remains. I have multiple Roku devices from [Roku 3’s and 4’s] and this issue happens on all of them. I do not think it is a Roku issue.
When the episode stops I can go back into the episode again fast forward to before or after the 2 minute mark and it will continue to play until the end. It also only happens on the first episode I am playing during a viewing session. I select the episode, doesn’t seem to matter if it is at the beginning or stopped in the middle. It will stop playing at the 2 minute mark, I can proceed to the next episode and it does not happen,
If there is a setting that controls this, say to avoid the credits or something, I don’t see one on the Roku or in my TV library settings.
My server runs off a Q-Nap server if that makes a difference, also installed the latest firmware on that today and issue remains.
I am having the same issue on my Roku devices…TV episodes or movies end 3-5 minutes before the end. If I go back and fast forward to the end, the content is there and works fine. It only seems to happen during extended playback. Would be interested to hear from the Plex team if this is a Roku issue or a Plex issue.
I can confirm that this behavior, which was very rare for me up until a few months ago, increased significantly on my Roku Stick+ with recent builds and seems to occur almost every time on server build 1.18.6.2368. I do not notice it on my TCL Roku TV very often but the Stick+ is my daily driver so that observation could be confirmation bias. It has never happened on my BRAVIA Android TV or SHIELD.
My life being as busy as it is, I am never in a position to capture server and device logs to share with Plex team, but considering this issue seems so widespread, I’d hope they were able to reproduce it themselves.
Feb 20, 2020 17:22:52.852 [2484] DEBUG - Auth: authenticated user 1 as RandomResponseUnit
Feb 20, 2020 17:22:52.852 [0280] DEBUG - Request: [192.168.0.61:40458 (Allowed Network (Subnet))] GET /video/:/transcode/universal/session/03cd7821-1593-475e-814e-14a05822d5ec-202/base/01876.ts?protocol=hls&mediaIndex=0&waitForSegments=1 (6 live) TLS Signed-in Token (RandomResponseUnit)
Feb 20, 2020 17:22:52.852 [0280] DEBUG - Asked for segment 1876 from session.
Feb 20, 2020 17:22:52.852 [0280] WARN - Transcode runner appears to have died.
Feb 20, 2020 17:22:52.954 [0280] WARN - Timed out waiting for segment.
Feb 20, 2020 17:22:52.956 [1196] DEBUG - Completed: [192.168.0.61:40458] 404 GET /video/:/transcode/universal/session/03cd7821-1593-475e-814e-14a05822d5ec-202/base/01876.ts?protocol=hls&mediaIndex=0&waitForSegments=1 (6 live) TLS 103ms 288 bytes (pipelined: 370)
at which point it appears you started the show over from the beginning, scrobbled to the spot you originally were, and it starts having the same issue and starts sending empty segments:
Feb 20, 2020 17:24:07.499 [2484] DEBUG - Auth: authenticated user 1 as RandomResponseUnit
Feb 20, 2020 17:24:07.499 [0280] DEBUG - Request: [192.168.0.61:40586 (Allowed Network (Subnet))] GET /video/:/transcode/universal/session/03cd7821-1593-475e-814e-14a05822d5ec-202/base/02954.ts?protocol=hls&mediaIndex=0&waitForSegments=1 (12 live) TLS Signed-in Token (RandomResponseUnit)
Feb 20, 2020 17:24:07.499 [0280] DEBUG - Asked for segment 2954 from session.
Feb 20, 2020 17:24:07.499 [0280] WARN - Transcode runner appears to have died.
Feb 20, 2020 17:24:07.600 [0280] DEBUG - Sending back blank segment for 2954, we overestimated the number of segments.
Feb 20, 2020 17:24:07.601 [0280] DEBUG - Returning segment 2954 from session
Feb 20, 2020 17:24:07.601 [0280] DEBUG - Content-Length of C:\Program Files (x86)\Plex\Plex Media Server\Resources\empty.ts is 188 (of total: 188).
Feb 20, 2020 17:24:07.602 [1196] DEBUG - Completed: [192.168.0.61:40586] 200 GET /video/:/transcode/universal/session/03cd7821-1593-475e-814e-14a05822d5ec-202/base/02954.ts?protocol=hls&mediaIndex=0&waitForSegments=1 (12 live) TLS 102ms 188 bytes (pipelined: 25)
The transcoding statistics files show that it transcoded the whole video however. And when it finished (which is just gives in milliseconds, yay math) thats the exact second the transcode runner dies and 8 seconds later youre trying to play that episode again. So I assume thats when the stream cut out. @ChuckPa Sorry to invoke the summoning ritual, but youre the best guy I know for this stuff. Any ideas?
So it seems that when the transcoder stops and doesnt send empty segments is when the stream cuts out. Then it times out waiting for segments. But when it does send empty segments, the show keeps playing.
Also a weird part in the spot where it died. I’m assuming you didnt seek for a new spot in the episode .7 seconds before the stream cut out.
Feb 20, 2020 17:22:53.682 [4644] DEBUG - Asked for segment 1877 from session.
Feb 20, 2020 17:22:53.682 [4644] WARN - Transcode runner appears to have died.
Feb 20, 2020 17:22:53.682 [5680] DEBUG - Streaming Resource: Resetting play progress due to seek
Feb 20, 2020 17:22:53.784 [4644] WARN - Timed out waiting for segment.
Feb 20, 2020 17:22:53.784 [1196] DEBUG - Completed: [192.168.0.61:40567] 404 GET /video/:/transcode/universal/session/03cd7821-1593-475e-814e-14a05822d5ec-202/base/01877.ts?protocol=hls&mediaIndex=0&waitForSegments=1 (6 live) TLS 103ms 288 bytes (pipelined: 1)
Thanks for looking at it. There are at least 4 other recent threads on the Roku section of the forum where people have what seems to be the same issue.
