WIP: Replace packet_debian9-macvlan with debian11
What type of PR is this?
/kind cleanup
What this PR does / why we need it:
packet_debian9-macvlan is unstable job. In addition, debian9 is going to EOL soon as 1 So this replaces packet_debian9-macvlan with debian11 to avoid wasting our time for the job anymore.
Which issue(s) this PR fixes:
Fixes #8916
Does this PR introduce a user-facing change?:
Drop debian9 support from Kubespray because of the EOL
[APPROVALNOTIFIER] This PR is APPROVED
This pull-request has been approved by: oomichi
The full list of commands accepted by this bot can be found here.
The pull request process is described here
- ~~OWNERS~~ [oomichi]
Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment
https://github.com/kubernetes-sigs/kubespray/blob/889454f2bc35dbfc69da8824d2e8bc658ba08412/README.md?plain=1#L120 We should remove Jessie and Stretch from README to clearly state that debian 9 (and 8 lol..) are not supported anymore
I think it's still worthwhile testing macvlan if we can get it to work but on more recent os like debian 11. WDYT?
I think it's still worthwhile testing macvlan if we can get it to work but on more recent os like debian 11. WDYT?
Indeed, maybe move the job to debian11 to see what's happening (not hopeful but well 😆)
I think it's still worthwhile testing macvlan if we can get it to work but on more recent os like debian 11. WDYT?
Indeed, maybe move the job to debian11 to see what's happening (not hopeful but well 😆)
From my investigation into the macvlan job, shared with @oomichi on slack just restarting containerd at the end of the run results in a working cluster so the issue is a race condition between when containerd is set up (first) and when the mtacvlan CNI is configured (second). We would need a generic way of restarting the container manager when the macvlan CNI is configured (I suspect this issue affects docker and cri-o as well, thus why "generic").
I think it's still worthwhile testing macvlan if we can get it to work but on more recent os like debian 11. WDYT?
Indeed, maybe move the job to debian11 to see what's happening (not hopeful but well 😆)
From my investigation into the macvlan job, shared with @oomichi on slack just restarting containerd at the end of the run results in a working cluster so the issue is a race condition between when containerd is set up (first) and when the mtacvlan CNI is configured (second). We would need a generic way of restarting the container manager when the macvlan CNI is configured (I suspect this issue affects docker and cri-o as well, thus why "generic").
handlers:
- include: container-engine/cri-o/handlers/main.yml
- include: container-engine/docker/handlers/main.yml
- include: container-engine/containerd/handlers/main.yml
- name: restart cri-o
command: /bin/true
notify: restart crio
when: container_manager == 'crio'
- name: restart containerd
command: /bin/true
notify: restart containerd
when: container_manager == 'containerd'
- name: restart docker
command: /bin/true
notify: restart docker
when: container_manager == 'docker'
😆 not sure it's the more elegant way though
I think it's still worthwhile testing macvlan if we can get it to work but on more recent os like debian 11. WDYT?
Indeed, maybe move the job to debian11 to see what's happening (not hopeful but well 😆)
Oh, that is a nice point. OK, let me try to convert the job to debian11 with macvlan. Let's see the test result.
Let's wake the bot.
/ok-to-test
@oomichi you will need to add the handlers suggested by @floryut for the job to work, even with the rebase to debian 11 the race condition is the same
@oomichi you will need to add the handlers suggested by @floryut for the job to work, even with the rebase to debian 11 the race condition is the same
I missed that. OK, let me try it.
/hold
Warning NetworkNotReady 38s (x151 over 5m38s) kubelet network is not ready: container runtime network not ready: NetworkReady=false reason:NetworkPluginNotReady message:Network plugin returns error: cni plugin not initialized
Try adding restarting-containerd
Even if restarting containerd,
" Type Reason Age From Message", " ---- ------ ---- ---- -------",
" Normal Scheduled 6m22s default-scheduler Successfully assigned default/netchecker-agent-dhdt7 to instance-2",
" Warning FailedCreatePodSandBox 6m22s kubelet Failed to create pod sandbox: rpc error: code = Unknown desc = failed to setup network for sandbox \"ba37310f2aa7f4311faf656bcce70f2851605cfd0481c927c98a3cd13478036a\": plugin type=\"macvlan\" name=\"mynet\" failed (add): Link not found",
" Warning FailedCreatePodSandBox 6m9s kubelet Failed to create pod sandbox: rpc error: code = Unknown desc = failed to setup network for sandbox \"dbd73beb158c2f715508f4f9c3ad2b943ae0e002f984bf17e9717ce2a7979180\": plugin type=\"macvlan\" name=\"mynet\" failed (add): Link not found",
" Warning FailedCreatePodSandBox 5m57s kubelet Failed to create pod sandbox: rpc error: code = Unknown desc = failed to setup network for sandbox \"e1881466321e89e65f426cba67e59b47d8581b773813e57053fb722be3448f93\": plugin type=\"macvlan\" name=\"mynet\" failed (add): Link not found",
" Warning FailedCreatePodSandBox 5m42s kubelet Failed to create pod sandbox: rpc error: code = Unknown desc = failed to setup network for sandbox \"f8ef25003c34a6a90605cdb674be1039e7ffa70a6db9f6a342a4bc40dd3f2fd1\": plugin type=\"macvlan\" name=\"mynet\" failed (add): Link not found",
" Warning FailedMount 5m37s kubelet MountVolume.SetUp failed for volume \"kube-api-access-2fzz4\" : failed to fetch token: Post \"https://localhost:6443/api/v1/namespaces/default/serviceaccounts/netchecker-agent/token\": EOF",
" Warning FailedMount 5m36s kubelet MountVolume.SetUp failed for volume \"kube-api-access-2fzz4\" : failed to fetch token: Post \"https://localhost:6443/api/v1/namespaces/default/serviceaccounts/netchecker-agent/token\": read tcp 127.0.0.1:34834->127.0.0.1:6443: read: connection reset by peer",
" Warning FailedMount 5m31s kubelet MountVolume.SetUp failed for volume \"kube-api-access-2fzz4\" : failed to fetch token: serviceaccounts \"netchecker-agent\" is forbidden: User \"system:node:instance-2\" cannot create resource \"serviceaccounts/token\" in API group \"\" in the namespace \"default\": no relationship found between node 'instance-2' and this object",
" Warning FailedCreatePodSandBox 5m29s kubelet Failed to create pod sandbox: rpc error: code = Unknown desc = failed to setup network for sandbox \"8b551a4fc07662e607fa85e9c2be9fc20ae12b9ad11b9aeb3071847ffc70c5f8\": plugin type=\"macvlan\" name=\"mynet\" failed (add): Link not found",
" Warning FailedCreatePodSandBox 5m17s kubelet Failed to create pod sandbox: rpc error: code = Unknown desc = failed to setup network for sandbox \"5d958c5ba58f0ee5cf06d35de8496fb82897c9f2fa2882f01ab8488a8d052f2f\": plugin type=\"macvlan\" name=\"mynet\" failed (add): Link not found",
" Warning FailedCreatePodSandBox 5m5s kubelet Failed to create pod sandbox: rpc error: code = Unknown desc = failed to setup network for sandbox \"0c54bcc5e1e7f2027bae719b6e8b246d7263ccf8cbfd6a14592205785309fb14\": plugin type=\"macvlan\" name=\"mynet\" failed (add): Link not found",
" Warning FailedCreatePodSandBox 4m54s kubelet Failed to create pod sandbox: rpc error: code = Unknown desc = failed to setup network for sandbox \"27fc30256a1fc42f196d7be289e13e8a4437ee00f053782f6a443175004f5f1f\": plugin type=\"macvlan\" name=\"mynet\" failed (add): Link not found",
" Warning FailedCreatePodSandBox 4m41s kubelet Failed to create pod sandbox: rpc error: code = Unknown desc = failed to setup network for sandbox \"085cf5b33c54e815db2a2b4bbe523ddf0d46155af3660e55a435d90a2fc13609\": plugin type=\"macvlan\" name=\"mynet\" failed (add): Link not found",
" Warning FailedCreatePodSandBox 4m27s kubelet Failed to create pod sandbox: rpc error: code = Unknown desc = failed to setup network for sandbox \"1744eb0b153322c836090fc6e88639507310da7e1a3b303852cd84e458e0c5d8\": plugin type=\"macvlan\" name=\"mynet\" failed (add): Link not found",
" Warning FailedCreatePodSandBox 4m15s kubelet Failed to create pod sandbox: rpc error: code = Unknown desc = failed to setup network for sandbox \"111ca9acfe017527ba7939ac8f9b11e521150986e384bc8d3207041d3cab9e3b\": plugin type=\"macvlan\" name=\"mynet\" failed (add): Link not found",
" Warning FailedCreatePodSandBox 4m2s kubelet Failed to create pod sandbox: rpc error: code = Unknown desc = failed to setup network for sandbox \"33c8cae75ae5fbcf80469f5747e1dba3160d1433a0b94ec0119d797621b2c1b6\": plugin type=\"macvlan\" name=\"mynet\" failed (add): Link not found",
" Warning FailedCreatePodSandBox 3m48s kubelet Failed to create pod sandbox: rpc error: code = Unknown des
[2890](https://gitlab.com/kargo-ci/kubernetes-sigs-kubespray/-/jobs/3051982087#L2890)c = failed to setup network for sandbox \"59a51a447c178183abdab96493dd56d30604ae8f567d904aab3560eec2ceb573\": plugin type=\"macvlan\" name=\"mynet\" failed (add): Link not found",
" Warning FailedCreatePodSandBox 18s (x16 over 3m34s) kubelet (combined from similar events): Failed to create pod sandbox: rpc error: code = Unknown desc = failed to setup network for sandbox \"f8e661abe843b6042285656bee66221da80841c603c6b6bbcd49a83943424bda\": plugin type=\"macvlan\" name=\"mynet\" failed (add): Link not found"]}
The Kubernetes project currently lacks enough contributors to adequately respond to all issues and PRs.
This bot triages issues and PRs according to the following rules:
- After 90d of inactivity,
lifecycle/staleis applied - After 30d of inactivity since
lifecycle/stalewas applied,lifecycle/rottenis applied - After 30d of inactivity since
lifecycle/rottenwas applied, the issue is closed
You can:
- Mark this issue or PR as fresh with
/remove-lifecycle stale - Mark this issue or PR as rotten with
/lifecycle rotten - Close this issue or PR with
/close - Offer to help out with Issue Triage
Please send feedback to sig-contributor-experience at kubernetes/community.
/lifecycle stale
The Kubernetes project currently lacks enough active contributors to adequately respond to all issues and PRs.
This bot triages issues and PRs according to the following rules:
- After 90d of inactivity,
lifecycle/staleis applied - After 30d of inactivity since
lifecycle/stalewas applied,lifecycle/rottenis applied - After 30d of inactivity since
lifecycle/rottenwas applied, the issue is closed
You can:
- Mark this issue or PR as fresh with
/remove-lifecycle rotten - Close this issue or PR with
/close - Offer to help out with Issue Triage
Please send feedback to sig-contributor-experience at kubernetes/community.
/lifecycle rotten
The Kubernetes project currently lacks enough active contributors to adequately respond to all issues and PRs.
This bot triages PRs according to the following rules:
- After 90d of inactivity,
lifecycle/staleis applied - After 30d of inactivity since
lifecycle/stalewas applied,lifecycle/rottenis applied - After 30d of inactivity since
lifecycle/rottenwas applied, the PR is closed
You can:
- Reopen this PR with
/reopen - Mark this PR as fresh with
/remove-lifecycle rotten - Offer to help out with Issue Triage
Please send feedback to sig-contributor-experience at kubernetes/community.
/close
@k8s-triage-robot: Closed this PR.
In response to this:
The Kubernetes project currently lacks enough active contributors to adequately respond to all issues and PRs.
This bot triages PRs according to the following rules:
- After 90d of inactivity,
lifecycle/staleis applied- After 30d of inactivity since
lifecycle/stalewas applied,lifecycle/rottenis applied- After 30d of inactivity since
lifecycle/rottenwas applied, the PR is closedYou can:
- Reopen this PR with
/reopen- Mark this PR as fresh with
/remove-lifecycle rotten- Offer to help out with Issue Triage
Please send feedback to sig-contributor-experience at kubernetes/community.
/close
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.