New Flow Plugin - Pause Node and Requeue

[davidry]

Feature Request

Occasionally, one of my nodes looses the ability to perform Hardware transcoding, I currently use CheckNodeHardwareEncoder Currently, I send the file to "Require Review", however this will fill-up my staging queue preventing the others from running

Ideally, Having a plugin that allows us to Pause a specific Node, and send the task to another worker would allow the service to continue working.

Bonus points:

• Allow Shutdown of a docker node
• Reboot worker / Reboot host for Non-Dockerized deployments
• Pause node for XX Minutes / One off Schedule

Note: I'm using Docker Swarm with a secondary stack to version: '3.8' services: DeviceMappingMgr: image: alpinelinux/docker-cli deploy: mode: global entrypoint: docker command: | run --rm -i --name device-manager --privileged --cgroupns=host --pid=host --userns=host -v /sys:/host/sys -v /var/run/docker.sock:/var/run/docker.sock ghcr.io/allfro/allfro/device-mapping-manager:nightly volumes: - /var/run/docker.sock:/var/run/docker.sock

[HaveAGitGat]

Maybe this well help, you can try the requeue system here:

https://github.com/HaveAGitGat/Tdarr/issues/748#issuecomment-2146710875

[HaveAGitGat]

2.20.01 released which can use Requeue and Worker Type flow plugins.

[davidry]

Thanks for that!

@HaveAGitGat - any chance we can have a plugin to pause the current node (Instead of Pause All Nodes), or Remove a worker-type tag? an alternate would be to send a shutdown worker request, forcing docker to re-start the worker

All my Docker instances start as VAAPI accelerated, however after some host-side process causes High CPU usage, the device claims to not be Hardware accelerated when running "check node hardware encoder" until I restart the container

[HaveAGitGat]

Removing the tag would be the same as setting it to CPUorGPU which allows any worker type to do it:

Or do you mean something different?

And sure will try to add a pause node plugin soon.

[davidry]

for some reason, a VAAPI compatible device will stop being able to utilize hardware encoding until the docker instance is rebooted:

this worker has been running for a few days, and has recently gotten stuck

having the ability to disable or un-mark the node's hardware type would prevent this loop Better yet, being able to shutdown the node would allow docker to re-create the worker and continue.

I'm not sure what happens to cause this, these services could run for days without issues, or run into this within the first few hours this has happened to multiple nodes.