-
Notifications
You must be signed in to change notification settings - Fork 315
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
[EKS] [request]: Kubernetes 1.13 Patch Update (to 1.13.8) #411
Comments
As far as I can tell, this only affects the kubelet, right? So it would be sufficient to use version 1.13.8 of the kubelet to get the fix. |
I confirm that rebuilding the ami using kubelet v1.13.8 fixed the issue. |
@mogren Now that the control plane is on 1.13.8 since last weekend, would you please be able to release a new AMI with 1.13.8? |
FYI: the 1.13.8 AMI suddenly changed docker group id from 995 to 994, so you need to adjust your build scripts and Dockerfiles accordingly. Tail of the
|
@pawelprazak |
Yes, not relying on the |
Looks like they need to get patches out for control-plane and worker images for security updates 1.13.10 in this case |
@aws would it be possible for you guys to publish eks node ami's as soon as patches come out? In case of 1.13.8 our autoscaling is not working as expected and in 1.13.10 it is clear that dos attacks can be a threat. Seems that we've to wait an additional month or two, which imho beats the purpose of patches. |
Closing this as EKS no longer supports Kubernetes v1.13. |
Tell us about your request
Thanks for supporting kubernetes v1.13 on EKS!
This issue is to see if EKS can bump the Kubernetes patch version from v1.13.7 to v1.13.8.
Which service(s) is this request for?
EKS - both the control plane and eks-optimized AMIs
Tell us about the problem you're trying to solve. What are you trying to do, and why is it hard?
Kubernetes 1.13.8 has an important fix (it fixes a regression that was introduced in v1.13) - especially pertaining to EKS' use of Amazon's VPC CNI plugin: when IP addresses are in the cooling period (30 seconds after pod termination when those IPs can't be used), this patch fixes the regression where kubelet no longer retries the sandbox creation on failure (regardless of the restart policy of the pod).
Are you currently working around this issue?
Until this is live, we have to revert to using EKS' Kubernetes 1.12 (which is less than ideal). Thanks for any updates here!
The text was updated successfully, but these errors were encountered: