Same. Running Version 1.41.3.9314 and all seems stable again. Will watch this thread for updates on when (if wanted) we can switch back to beta channel. That’s what beta is for right? 
1 Like
What used to consume 50GB of space in a matter of minutes, I’m now at 10GB after 20 minutes and slowly climbing.
1 Like
Working well for me too. When the RAM fills up it actually purges older data chunks so nothing crashes.
2 Likes
Everyone happy with the changes ?
2 Likes
So far so good, yes. I was able to confirm that data does get purged.
2 Likes
back to normal, yes. Regardless of how fast new GPU tech is, the throttle function is still desired, for a few reasons. I’m not running my GPU at 100% for the whole video, if 10 streams start in a relatively short time in order 1, 2, then 3 they don’t progressively get slower to start with throttling. Users often start a video, if they don’t like it they move on, now I have wasted GPU time transcoding/writing data that will never get read. I skimmed over this discussion, but I assume the attempt was to remove this feature?, It DOES work, I see my GPU go from 100% duty cycle to small peaks of transcoding when throttling. if the worry is it goes into slow transcode mode vs stop, I don’t follow the issue? If a user closes the player it stops right? Or maybe if paused for more than 1 hour it should also stop, they need to restart transcode when they un-pause. a few ways to skin a cat.
I am running 1.41.4.9463 and the issue appears to be resolved. Thanks for everyone’s helpful troubleshooting on this.
1 Like
I’m on 1.41.4.9463 and just started experiencing this issue daily. Is there another fix I’m missing?
I’m on 1.41.5.9522 and see transcoding filling up my 8GB /dev/shm.
Usually around 40-45 min into a 4K to 1080P transcode with tonemapping enabled, hw transcoding with an NVIDIA card.
EDIT: I am watching the session folder and it stopped cleaning up since “media-00555.ts” on a 120 second transcode throttle buffer setting.
EDIT2: It eventually cleans up older files but progressively gets slower at it, resulting in /dev/shm becoming full.
This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.
