error execution phase kubelet-start: error uploading crisocket: timed out waiting for the condition

[kcalmond]

Details

What steps did you take and what happened:

Spun up a 4 node cluster (3 masters, 1 worker). His this error when reaching TASK [cluster : join | add node to cluster]:

TASK [cluster : join | add node to cluster] ******************************************************************************************************************************************
Sunday 14 February 2021  12:17:44 -0800 (0:00:00.919)       0:09:29.725 *******
changed: [blackberry]
fatal: [strawberry]: FAILED! => changed=true
  cmd:
  - kubeadm
  - join
  - --config
  - /etc/kubernetes/kubeadm-join.yaml
  delta: '0:04:12.506743'
  end: '2021-02-14 20:21:57.663006'
  msg: non-zero return code
  rc: 1
  start: '2021-02-14 20:17:45.156263'
  stderr: |2-
            [WARNING SystemVerification]: missing optional cgroups: hugetlb
    error execution phase kubelet-start: error uploading crisocket: timed out waiting for the condition
    To see the stack trace of this error execute with --v=5 or higher
  stderr_lines: <omitted>

More details in this comment: #107 (comment)

What did you expect to happen:
Successful execution of all plays resulting in a three-master HA cluster with one worker node.

Anything else you would like to add:

Worked around the error to get a healthy cluster. Process:

• edit inventory so only the offending master node "strawberry" is included (just commented all the other node lines out).
• ran the nuke play: > env ANSIBLE_CONFIG=ansible/ansible.cfg ansible-playbook ansible/playbooks/nuke.yml
• Still needed to remove the node from the cluster:

> kubectl get nodes --kubeconfig ansible/playbooks/output/k8s-config.yaml
NAME         STATUS   ROLES                  AGE     VERSION
blackberry   Ready    <none>                 24h     v1.20.2
blueberry    Ready    control-plane,master   24h     v1.20.2
hackberry    Ready    control-plane,master   24h     v1.20.2
strawberry   NotReady <none>                 8m24s   v1.20.2
 
> kubectl delete node strawberry --kubeconfig ansible/playbooks/output/k8s-config.yaml
node "strawberry" deleted

• Then - restored inventory to include all nodes ref
• Then ran the all.yml play again, which successfully added "strawberry" as a third master node.

[fixxar1911]

I had this same issue!