Add image_arch in flannel image tag

[muzi502]

What type of PR is this?

/kind bug

What this PR does / why we need it:

#6166 enabled multi-CPU architecture in Flannel CNI

For flannel image, in download image task flannel image tag is flannel_version, such as quay.io/coreos/flannel:v0.13.0 but in roles/network_plugin/flannel/templates/cni-flannel.yml.j2 file, real image tag is {{ flannel_image_tag }}-{{ arch }}, such as quay.io/coreos/flannel:v0.13.0-amd64 .

This will result in the real image tag is inconsistent with the downloaded image tag especially for offline deploy.

• this is download flannel image task log

hub.k8s.li is my private docker registry

TASK [download : check_pull_required | Set pull_required if the desired image is not yet loaded] *********************************************************************************
ok: [kube-control-3]
ok: [kube-control-2]
ok: [kube-control-1]
ok: [kube-node-1]
Tuesday 27 April 2021  09:44:51 +0000 (0:00:00.160)       0:02:03.313 *********
Tuesday 27 April 2021  09:44:51 +0000 (0:00:00.119)       0:02:03.432 *********

TASK [download : debug] **********************************************************************************************************************************************************
ok: [kube-control-3] => {
    "msg": "Pull hub.k8s.li/coreos/flannel:v0.13.0 required is: False"
}
ok: [kube-control-2] => {
    "msg": "Pull hub.k8s.li/coreos/flannel:v0.13.0 required is: False"
}
ok: [kube-control-1] => {
    "msg": "Pull hub.k8s.li/coreos/flannel:v0.13.0 required is: False"
}
ok: [kube-node-1] => {
    "msg": "Pull hub.k8s.li/coreos/flannel:v0.13.0 required is: False"
}

• this is kube-flannel status when complete deploy

kube-system   kube-flannel-47krk                       0/1     Init:ImagePullBackOff   0          48m
kube-system   kube-flannel-cqcjc                       0/1     Init:ImagePullBackOff   0          48m
kube-system   kube-flannel-flpcz                       0/1     Init:ImagePullBackOff   0          48m
kube-system   kube-flannel-xh2pt                       0/1     Init:ImagePullBackOff   0          48m
  Normal   Scheduled  80s                default-scheduler  Successfully assigned kube-system/kube-flannel-lsjxk to kube-control-1
  Normal   Pulling    37s (x3 over 80s)  kubelet            Pulling image "hub.k8s.li/coreos/flannel:v0.13.0-amd64"
  Warning  Failed     34s (x3 over 77s)  kubelet            Failed to pull image "hub.k8s.li/coreos/flannel:v0.13.0-amd64": rpc error: code = NotFound desc = failed to pull and unpack image "hub.k8s.li/coreos/flannel:v0.13.0-amd64": failed to resolve reference "hub.k8s.li/coreos/flannel:v0.13.0-amd64": hub.k8s.li/coreos/flannel:v0.13.0-amd64: not found
  Warning  Failed     34s (x3 over 77s)  kubelet            Error: ErrImagePull
  Normal   BackOff    8s (x4 over 77s)   kubelet            Back-off pulling image "hub.k8s.li/coreos/flannel:v0.13.0-amd64"
  Warning  Failed     8s (x4 over 77s)   kubelet            Error: ImagePullBackOff

Which issue(s) this PR fixes:

Fixes #

Special notes for your reviewer:

Does this PR introduce a user-facing change?:

[k8s-ci-robot]

Hi @muzi502. Thanks for your PR.

I'm waiting for a kubernetes-sigs member to verify that this patch is reasonable to test. If it is, they should reply with /ok-to-test on its own line. Until that is done, I will not automatically test new commits in this PR, but the usual testing commands by org members will still work. Regular contributors should join the org to skip this step.

Once the patch is verified, the new status will be reflected by the ok-to-test label.

I understand the commands that are listed here.

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.

[floryut]

Nice, shouldn't this also have image arch 🤔

kubespray/roles/network_plugin/canal/templates/canal-node.yaml.j2

Line 84 in 0feec14

image: "{{ flannel_image_repo }}:{{ flannel_image_tag }}"

[muzi502]

Nice, shouldn't this also have image arch 🤔

kubespray/roles/network_plugin/canal/templates/canal-node.yaml.j2

Line 84 in 0feec14

image: "{{ flannel_image_repo }}:{{ flannel_image_tag }}"

Sorry i forgot canal 😅

should remove -{{ arch }} and add it in flannel_image_tag , such as flannel_image_tag: "{{ flannel_version }}-{{ image_arch }}" 🤔️

kubespray/roles/network_plugin/flannel/templates/cni-flannel.yml.j2

Line 69 in 0feec14

image: {{ flannel_image_repo }}:{{ flannel_image_tag }}-{{ arch }}

kubespray/roles/download/defaults/main.yml

Line 404 in 0feec14

flannel_image_tag: "{{ flannel_version }}"

Other images set image arch in it's image_repo or image_tag, such as dnsautoscaler and calico, only flannel set image arch in its's template file.

kubespray/roles/download/defaults/main.yml

Line 459 in 0feec14

dnsautoscaler_image_repo: "{{ kube_image_repo }}/cpa/cluster-proportional-autoscaler-{{ image_arch }}"

kubespray/roles/download/defaults/main.yml

Line 406 in 0feec14

calico_node_image_tag: "{{ calico_version }}{%- if image_arch != 'amd64' -%}-{{ image_arch }}{%- endif -%}"

find roles -type f -name "*.j2" | xargs grep '_tag }}' | grep arch
roles/network_plugin/flannel/templates/cni-flannel.yml.j2:        image: {{ flannel_image_repo }}:{{ flannel_image_tag }}-{{ arch }}
roles/network_plugin/flannel/templates/cni-flannel.yml.j2:        image: {{ flannel_image_repo }}:{{ flannel_image_tag }}-{{ arch }}

[muzi502]

/cc @floryut
I add image_arch in flannel_image_tag, then use regex_replace(image_arch,'') to remove redundant image_arch

https://github.com/kubernetes-sigs/kubespray/blob/10083fa47e579b49f7cd9736556bd9390e1ecb92/roles/download/defaults/main.yml#L404

kubespray/roles/network_plugin/flannel/templates/cni-flannel.yml.j2

Line 69 in 3ed0b6a

image: {{ flannel_image_repo }}:{{ flannel_image_tag | regex_replace(image_arch,'') }}{{ arch }}

flannel's task will create five daemonset with different arch, should keep {{ arch }} variable

[root@kube-control-1 ~]# kubectl get daemonset -n kube-system
NAME                                     DESIRED   CURRENT   READY   UP-TO-DATE   AVAILABLE   NODE SELECTOR            AGE
daemonset.apps/kube-flannel              4         4         4       0            4           <none>                   73m
daemonset.apps/kube-flannel-ds-arm       0         0         0       0            0           <none>                   73m
daemonset.apps/kube-flannel-ds-arm64     0         0         0       0            0           <none>                   73m
daemonset.apps/kube-flannel-ds-ppc64le   0         0         0       0            0           <none>                   73m
daemonset.apps/kube-flannel-ds-s390x     0         0         0       0            0           <none>                   73m

This will keep download flannel image tag and real used image tag consistent

[muzi502]

/kind bug

[oomichi]

/ok-to-test

[oomichi]

I was wondering why we need to remove and add similar string here, but that seems the same as the existing code https://github.com/kubernetes-sigs/kubespray/blob/10083fa47e579b49f7cd9736556bd9390e1ecb92/roles/network_plugin/flannel/templates/cni-flannel.yml.j2#L69

So it is fine to do that.

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: floryut, muzi502

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [floryut]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[oomichi]

/lgtm