Set kubelet's cgroup driver to groupfs in playbook

[hangyan]

In v1.22, if the user is not setting the cgroupDriver field under
KubeletConfiguration, kubeadm will default it to systemd.

Signed-off-by: Hang Yan yhang@vmware.com

[hangyan]

/assign lzhecheng

[codecov-commenter]

Codecov Report

Merging #2883 (6927341) into main (4d0eea7) will decrease coverage by 1.79%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##             main    #2883      +/-   ##
==========================================
- Coverage   60.30%   58.50%   -1.80%     
==========================================
  Files         283      283              
  Lines       23504    23504              
==========================================
- Hits        14173    13750     -423     
- Misses       7797     8275     +478     
+ Partials     1534     1479      -55     

Flag	Coverage Δ
kind-e2e-tests	45.21% <ø> (-2.58%)	⬇️
unit-tests	40.91% <ø> (-0.02%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
...g/agent/apiserver/handlers/featuregates/handler.go	0.00% <0.00%> (-82.36%)	⬇️
...kg/apiserver/registry/system/supportbundle/rest.go	19.54% <0.00%> (-55.18%)	⬇️
pkg/support/dump_others.go	0.00% <0.00%> (-51.73%)	⬇️
...ver/registry/controlplane/nodestatssummary/rest.go	50.00% <0.00%> (-50.00%)	⬇️
pkg/support/dump.go	8.13% <0.00%> (-49.60%)	⬇️
...g/agent/apiserver/handlers/addressgroup/handler.go	0.00% <0.00%> (-40.00%)	⬇️
...agent/apiserver/handlers/appliedtogroup/handler.go	0.00% <0.00%> (-40.00%)	⬇️
pkg/apiserver/handlers/loglevel/handler.go	0.00% <0.00%> (-38.47%)	⬇️
pkg/agent/controller/networkpolicy/packetin.go	46.00% <0.00%> (-26.00%)	⬇️
pkg/ovs/ovsctl/ofctl.go	18.51% <0.00%> (-19.76%)	⬇️
... and 23 more

[antoninbas]

Could you clarify why this change is needed:

• what breaks without this change?
• the K8s documentation recommends using systemd, so why don't we try to get systemd to work instead?

[hangyan]

Could you clarify why this change is needed:

* what breaks without this change?

* the K8s documentation recommends using systemd, so why don't we try to get systemd to work instead?

Sorry my comment was not clear before.

1. docker is using cgroupfs by default. In v1.22, since kubelet change to systemd, it won't run by default. They need to match.
2. The other option is change docker's to systemd too. But i don't find many resources on whether or why which one is better. Since the playbook's target is provding a testenv, i'm assuming it won't cause any big influence on our daily test(it works fine before v1.22)

[antoninbas]

@hangyan I think that's in part where my confusion comes from. The playbook has already been updated a couple of months ago to use containerd instead of docker: #2583. So we are not using docker anymore.

And before we switched to containerd, we had already updated the playbook to make docker use systemd: #2543

[hangyan]

Sorry i miss that. I was using a downstream project which contains ref to the old antrea version which is before #2543. Thanks for the explaination. Will update downstream code and close this PR now.