[Feature] transport layer offload optimisation #975

Closed
opened 2025-12-29 02:26:58 +01:00 by adam · 2 comments
Owner

Originally created by @zilexa on GitHub (Mar 17, 2025).

Use case

This is a question in the first place.

Please see here the explanation:
https://tailscale.com/kb/1320/performance-best-practices#ethtool-configuration

I was wondering if the docker image already takes care of this, or if we need to configure this on the host machine?

Description

For example, for Fedora Silverblue/uBlue OS/Bluefin/Bazzite as host OS, this command is required to run once to enable this configuration for Tailscale running directly on the host:

https://discussion.fedoraproject.org/t/tailscale-with-transport-layer-offload-optimisation/146995/2

Contribution

  • I can write the design doc for this feature
  • I can contribute this feature

How can it be implemented?

No response

Originally created by @zilexa on GitHub (Mar 17, 2025). ### Use case This is a question in the first place. Please see here the explanation: https://tailscale.com/kb/1320/performance-best-practices#ethtool-configuration I was wondering if the docker image already takes care of this, or if we need to configure this on the host machine? ### Description For example, for Fedora Silverblue/uBlue OS/Bluefin/Bazzite as host OS, this command is required to run once to enable this configuration for Tailscale running directly on the host: https://discussion.fedoraproject.org/t/tailscale-with-transport-layer-offload-optimisation/146995/2 ### Contribution - [ ] I can write the design doc for this feature - [ ] I can contribute this feature ### How can it be implemented? _No response_
adam added the enhancement label 2025-12-29 02:26:58 +01:00
adam closed this issue 2025-12-29 02:26:58 +01:00
Author
Owner

@ArcticLampyrid commented on GitHub (Apr 4, 2025):

Headscale is a control-plane application and does not handle data forwarding (perhaps the built-in DERP server is an exception?), so there likely isn't much room for additional optimization.

@ArcticLampyrid commented on GitHub (Apr 4, 2025): Headscale is a control-plane application and does not handle data forwarding (perhaps the built-in DERP server is an exception?), so there likely isn't much room for additional optimization.
Author
Owner

@nblock commented on GitHub (Apr 5, 2025):

This is an optimization for the client that depends on the available kernel version and needs to be configured there.

@nblock commented on GitHub (Apr 5, 2025): This is an optimization for the client that depends on the available kernel version and needs to be configured there.
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: starred/headscale#975