Live TV Docker swarm crash

[prox32 2]

Hi all,

My set up is a 3 node docker swarm cluster managed with portainer.

My issue is when watching live TV for an extended period of time 20 minutes or so causes the worker node the container is on to fall over completely crashing the docker process but not the entire server, the actual tin stays up and working with no obvious source of the issue when monitoring while the issue occurs.

My first thought was resources on the box but generally the RAM usage is under 2GB and the cpu is 0 - 10%, this obviously increases while watching TV but the CPU only hits up to 60%, the RAM usage climbs steadily until around 2.1GB but doesn't seem to pass this out of 8GB total. I first thought it could be related to my load balancer so bypassed this to go directly to my docker cluster and still I see the same issue. This only happens when watching live TV on emby and not while watching other VOD content, I am not expecting an immediate solution but would appreciate any help in tracking down where this issue may be?

Edited September 15, 2022 by prox32

[roaku 842]

Is your transcode temp directory filling up or otherwise a bottleneck?

[prox32 2]

31 minutes ago, roaku said:

Is your transcode temp directory filling up or otherwise a bottleneck?

Well that was easy! Haha I changed this directory to a drive with multi terabytes free and haven't witnessed this behaviour so far.

Thank you for the extremely quick suggestion.