Lock kube-router version

[igorvpcleao]

Hi everyone,

It seems kops doesn't lock kube-router version. It may cause several problems as it may update to a newer version without user knowledge, and break things inexplicably.

On the template, it has only the image name, without a explicit version. That downloads latest, which may change and break:

kops/upup/models/cloudup/resources/addons/networking.kuberouter/k8s-1.6.yaml.template

Line 40 in 98b2437

image: cloudnativelabs/kube-router

This affected one of our clusters and took us quite some time to debug, as we could not pinpoint what happened over the night that caused connection errors. It turns out kube-router had a new release and some of our nodes were reset, so they were updated.

[andrewsykim]

@igorvpcleao @Cisneiros @tonidas is there a version of kube-router you know works very well with the latest version of kops?

[Cisneiros]

@andrewsykim We have been running 0.0.20 for about 17 hours straight with no errors.

[andrewsykim]

cc @murali-reddy

[andrewsykim]

If someone can open a PR for this, I can approve it :)

[chrislovecnm]

/assign @murali-reddy

[k8s-ci-robot]

@chrislovecnm: GitHub didn't allow me to assign the following users: murali-reddy.

Note that only kubernetes members can be assigned.

In response to this:

/assign @murali-reddy

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.

[BobbyJohansen]

A lock in on Kops create cluster would be a great option. or atleast a change on kops edit to specify the image to v0.0.20

[murali-reddy]

If any one can tell me what is the expected change i can submit a PR.

     image: cloudnativelabs/kube-router
     imagePullPolicy: Always

I can atlease change the pull policy to IfNotPresent. But i am not sure how do i go about the image version. What are other supported CNI plugins KOPS practice?

Should i lock it to latest stable version and update manifest once in a while (by sumbitting PR) to latest stable?

[BobbyJohansen]

I like either locking it or adding a variable to the template

kops/upup/models/cloudup/resources/addons/networking.kuberouter/k8s-1.6.yaml.template

Line 40 in 98b2437

image: cloudnativelabs/kube-router

to use a variable that comes from the kops cluster.spec section:

networking: kuberouter: {}

[chrislovecnm]

Should i lock it to latest stable version and update manifest once in a while (by sumbitting PR) to latest stable?

@murali-reddy it is your call on what version you set the container to, but we need a version number.

[andrewsykim]

#4935 locks the version, I will follow up later to propagate the kube router version from cluster spec

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale