The nVidia Team would like to see new recent evidence of the issue with hardware acceleration that was leading to low memory on the Shield and the Transcoder process getting killed and also in some cases the Plex Media Server process gets closed down as well.
Please see
https://support.plex.tv/articles/201643703-reporting-issues-with-plex-media-server/
https://support.plex.tv/articles/200250417-plex-media-server-log-files/
The nVidia Team would like to see new recent evidence of the issue with hardware acceleration that was leading to low memory on the Shield and the Transcoder process getting killed and also in some cases the Plex Media Server process gets closed down as well.
Please see
https://support.plex.tv/articles/201643703-reporting-issues-with-plex-media-server/
https://support.plex.tv/articles/200250417-plex-media-server-log-files/
Plex Transcoder Statistics.1.log (2.43 KB)
Plex Transcoder Statistics.log (2.29 KB)
Plex Media Server.log (232 KB)
(Attachment Plex on Nvidia 12-14-2014 Nvidia 7.2 Plex Version 1.14.1.5487.pptx is missing)
Thanks for the server logs.
I can see a conversion starting at 13:13:49 for /storage/emulated/0/TV Shows/Hawaii Five-0 (2010)/Season 09/Hawaii Five-0 (2010) - S09E10 - Pio Ke Kukui Po'\''ele Ka Hale.ts using hardware acceleration on the shield and then approx 3 and a half minutes later at 13:17:20, the process gets killed
It is possible that this was due to low memory. It can only be verified through a logcat using adb.
Could you provide me a sample that brings this on - preferably the actual file or the first 500Mb or so from the file and will try and reproduce the failure
Even the sample i provided a month ago still crashesā¦
I have added a power point with screen shots to show the process and failure
Plex on Nvidia 12-14-2014 Nvidia 7.2 Plex Version 1.14.1.5487.zip (292 KB)
I am not disputing that it does happen.
What I was after is having a reasonable size file that fails for you with HW Acceleration we can try and reproduce it and pass it to the nVidia team or the Plex Media Server development team as appropriate
The more evidence we have of the problem with different sample files and from different users the better understanding of the impact of the problem.
There is always a question mark that may be it is to do with the actual files that a user has - so having more sample perhaps from different source would be of benefit
i will try your original sample again
Today it is harder to find a file that actually works, most of them are failing after some time. I could provide dozen of files, the example is just in a reasonable size. As there are other users reporting the same, it can not relay on my installation. A user reported that even plex own PVR service is crashing like this - it is clearly not file related.
I am very disappointed that it takes so long to fix this critical bug.
Let me revert the question here, do you have a sample - big one - which doesnāt fail when you try to test with it ? if yes can you share it so that we also test and provide logs with your file?
Iāve been trying to reduce the recording time to get you the smallest file that failsā¦ In this example, I had a failure near the end of transcoding to āmobileā using a 15 minute 1080 file (server disconnected too). But I had success trying a second time trascoding the same file after the server re-connected. (I had first deleted all optimized versions so I could convert it a second time) I noticed that in both cases, the transcoder progress made it to over 100%, up to 104%. Probably unrelated. I believe that my fail.1.log documents what happened up until the disconnect, and fail.log is the reconnected log. The second trascoding should be reflected in the success log. I did some āsecond tryā conversions on some other longer files but had failures.
Plex Media Server Fail.1.log (4.2 MB)
Plex Media Server Fail.log (101.9 KB)
Plex Media Server success.log (411.2 KB)
Original File:
I can confirm that the issue is still arising when optimizing your 1.8Gb split2.mkv file
I was simply after more samples from other users
These may show something new. Unlike the 1080p examples, this show was recorded as 720 (H.264). The first couple of times I tried to transcode the recording, it failed (seen on the fail logs) Then I waited a couple of hours and tried again. This time, the transcoding failed at around the same point (80%) but did not disconnect, and did not show as a fail. It simply stared at zero again on itās own. On the successful transcode, before the system reset, Plex would transcode at a wide range of speeds, from 5.1% to 3.6% to 1:5%, back up the 4.7%, 5.0%, 2.3%, etc. After it restarted, it was a pretty solid 2.4% the entire time. When I went to another section of Plex and came back it was at 1.9, but got back to 2.4% immediately. I should note that I havenāt been doing anything but transcode during any of these tests. At the 2.4% speed, Plex finished the transcoding successfully. I donāt know if this is because the source was 720p, or if the system was throttling after the first attempt, or something else is at play, but on the second go round, I had success.
transcode 1 fail.zip (31.6 KB)
transcode 2 success.zip (47.2 KB)
Many samples here. Higher bitrate fail.
Below are a number of H.264 and H.265/HEVC encoded .mkv video clips that can be used for testing the network streaming and playback performance of media streamers & HTPCs.
Thank you for the diagnostics and the media file. I have managed to reproduce the failure using your supplied file. Please keep it available and i will pass the download link to our development team and the nVidia team to help with identifying the issue
I just want to chime in here. I am experiencing the same issues reported here. Nvidia Shield TV 2017 (16GB) running shield experience 7.2.1 and latest Plex Server.
Using hardware transcoding Plex Server on the Shield will restart after 10-20 minutes of transcoding. Disabling HW transcoding seems to fix the problem.
Reading about low memory I monitored memory usage through Kodi on the Shield while Plex Server on the Shield was performing HW transcoding. Free memory steadily decreased from about 1GB to under 500MB in a few minutes and Kodi started acting sluggish. Finally the transcoding failed and Plex Server restarted. Using software transcoding the memory usage stabilizes and free memory stays above 900MB.
I cannot remember having these problems before this summer and from the looks of it, low memory is triggering a kill of the Plex Server.
Iām having the same issue. Transcoding fails and crashes PMS after 5-10 minutes, this is a fault I encounter with transcoding for remote streaming and for local optimization.
Interestingly, I took a file that crashes during local transcoding every time and split it into 1GB chunks to post a sample here. As a whole the file will not transcode but each individual 1GB chunk will complete.
Adding myself to the list of afflicted. An unobtrusive, efficient unit with hardware transcoding is the reason I went with the Shield. From the above, it seem a fix is possible, it just hasnāt been accomplished yet (since August)?
Fortunately, Iām still within my return window for the Shield
Joining the list of affected users, Iām having the 2017 Shield Pro and having the same exact issue; the server crashes while transcoding and/or optimizing. Had to disable the hardware acceleration to let optimizations run, slowly but at least the server is not crashing.
Me two within the return window and would like to know if there is any progress on the issue?
Is there an update on the status of this issue? I am still experiencing transcoding problems at the 12-13 min mark to my Roku boxes. This problem has been going for months and has rendered my setup useless for Live TV/DVR when combined with the need for transcoding. That is the main reason I paid for my Lifetime Plex Pass and now cannot enjoy it.
Try 1.5Gb chunks. I always want recent examples - especially since there have been updates to kernel with the new NVIDIA 7.2.2 update
