upgrading to 1.3.2 but it remains 1.3.0

[pahud]

I was upgrading from 1.2.1 to 1.3.2 according to this document, however, it seems to be the cni version is 1.3.0.

amazon-vpc-cni-k8s/config/v1.3/aws-k8s-cni.yaml

Line 71 in 4a51ed4

- image: 602401143452.dkr.ecr.us-west-2.amazonaws.com/amazon-k8s-cni:1.3.0

pahud:~/environment $ kubectl describe daemonset aws-node --namespace kube-system | grep Image | cut -d "/" -f 2
amazon-k8s-cni:1.2.1
pahud:~/environment $ kubectl apply -f https://raw.githubusercontent.com/aws/amazon-vpc-cni-k8s/master/config/v1.3/aws-k8s-cni.yaml
clusterrole.rbac.authorization.k8s.io "aws-node" configured
serviceaccount "aws-node" unchanged
clusterrolebinding.rbac.authorization.k8s.io "aws-node" configured
daemonset.extensions "aws-node" configured
customresourcedefinition.apiextensions.k8s.io "eniconfigs.crd.k8s.amazonaws.com" configured
pahud:~/environment $ kubectl describe daemonset aws-node --namespace kube-system | grep Image | cut -d "/" -f 2
amazon-k8s-cni:1.3.0

How can I make sure it's already 1.3.2?

[pahud]

checking the log and it seems to remain v1.3.0

tail -f /var/log/aws-routed-eni/*

[bootc]

Indeed, https://github.com/aws/amazon-vpc-cni-k8s/blob/v1.3.2/config/v1.3/aws-k8s-cni.yaml is stuck on v1.3.0. If you change the tag to v1.3.2 then the image pull fails.

[sdavids13]

The latest config file on master has the 1.3.2 version set (permalink): https://raw.githubusercontent.com/aws/amazon-vpc-cni-k8s/0121d79566efa768275d3ec3d95fb7a599f84614/config/v1.3/aws-k8s-cni.yaml

The annoying part is that the version changed slightly to have a v prefixed so the image version is v1.3.2 instead of the old version of just 1.3.0.

Commit used to update the yaml file: 0121d79#diff-c5b46112cd953925bd7125f8ce8296e5

[Deshke]

docker pull 602401143452.dkr.ecr.us-west-2.amazonaws.com/amazon-k8s-cni:v1.3.2
Error response from daemon: Get https://602401143452.dkr.ecr.us-west-2.amazonaws.com/v2/amazon-k8s-cni/manifests/v1.3.2: no basic auth credentials

strangely enough both 1.3.0 and v1.3.2 now require some auth....

[mogren]

This was definitely a mistake in the 1.3 release branch. Put up a PR.

[tabern]

It looks like this was a previous commit on the 1.3 branch. This is updated in the latest master and 1.3 branch. https://github.com/aws/amazon-vpc-cni-k8s/blob/release-1.3/config/v1.3/aws-k8s-cni.yaml

[jkula-c]

Here is my solution:

1. login to aws ecr:
   aws ecr get-login --region us-west-2 --no-include-email

2. the above cmd returns somthing like this:
   sudo docker login -u AWS -p eyJwYXlsb2FkIjoiak5aRDNKdFVvZ3BtMjluaWJTMnlMdm1LREVXNndlc2dHK0kvL3ZnY1haS2NsS1ptUmJGYnRtNlNGL2tZMUhiTU1NdEZpN3NZM= https://602401143452.dkr.ecr.us-west-2.amazonaws.com

3. pull the image on my Local workstation:
   sudo docker pull 602401143452.dkr.ecr.us-west-2.amazonaws.com/amazon-k8s-cni:v1.3.2

4. Tag to my public repo:
   sudo docker tag 602401143452.dkr.ecr.us-west-2.amazonaws.com/amazon-k8s-cni:v1.3.2 Myusername/602401143452.dkr.ecr:v1.3.2

5. Push it
   sudo docker push Myusername/602401143452.dkr.ecr:v1.3.2

6. update the image on aws-k8s-cni.yaml

7. sudo kubectl -f apply