-
Notifications
You must be signed in to change notification settings - Fork 522
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
kubelet, containerd: replace 'containerd.sock' with 'dockershim.sock' #796
Conversation
I would like to see an issue opened to track:
Can either be |
23886a3
to
4e6ebb0
Compare
Addresses @bcressey 's comments. Removed the change from |
To confirm, this was done with the default |
Let containerd create the container runtime endpoint as '/run/dockershim.sock' instead of '/run/containerd/containerd.sock' Pass the new container runtime endpoint to kubelet as '/run/dockershim.sock'
4e6ebb0
to
91810c8
Compare
That is correct! |
Force push above rebases on develop |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Changes lgtm! All that's left now is to figure out how to undo 😎
e6523f0
to
ed00d96
Compare
Addresses @tjkirch 's comments. |
@bcressey @jahkeup @etungsten I know this is an old commit, but I'm wondering about the reasoning behind it. Was there a specific bug that changing the containerd sock path to something other than |
At the time, this was done for compatibility with the VPC CNI plugin, which didn't include the containerd socket in its mounts. It avoided the need for an edit to the CNI config as a prerequisite to getting started with Bottlerocket and EKS. We could likely switch back to the default path now, but would need to check how long the relevant fixes in the VPC CNI plugin have been deployed. |
Gotcha, OK that makes sense (and I vaguely remember the conversation in VPC CNI around this... :) ). I think it would be good to align back to the |
This changes containerd's container runtime socket to '/run/containerd/containerd.sock' from '/run/dockershim.sock'. This change reverts changes done in bottlerocket-os#796 More K8s deployments are now containerd-aware so customers are having to do daemonset edits for solutions that assume containerd as the underlying runtime. In the containerd systemd service, we create a symlink for the containerd socket at `/run/dockershim.sock` just so other deployments that assume docker as underlying runtime can still work.
Issue number: Fixes #747 (comment)
Description of changes:
Let containerd create the container runtime endpoint as
/run/dockershim.sock
instead of/run/containerd/containerd.sock
Pass the new container runtime endpoint to kubelet as
/run/dockershim.sock
Add changelog entry for v0.3.0
Testing done:
Built image, launched bottlerocket instance, connects to my cluster fine, pods can run fine.
Restarting
aws-node
pods and they still works fine.Checking
ipamd
shows IP addresses are being allocated successfully:Terms of contribution:
By submitting this pull request, I agree that this contribution is dual-licensed under the terms of both the Apache License, version 2.0, and the MIT license.