Upgrade node is failing with upgrade-cluster playbook

[kesslerdev]

Hello,

we have a bug when we try to upgrade a master node (indus5), nodes
indus3(master), indus11(worker), indus13(worker), indus15(worker) are already upgraded with success. we are using --limit to update only one node at time.

Error:
'ansible.vars.hostvars.HostVarsVars object' has no attribute 'kubeadm_init'

Environment:

• Cloud provider or hardware configuration:
  we are using Virtual machines provided by vmware ESX
• OS (printf "$(uname -srm)\n$(cat /etc/os-release)\n"):

Linux 3.10.0-957.1.3.el7.x86_64 x86_64
NAME="CentOS Linux"
VERSION="7 (Core)"
ID="centos"
ID_LIKE="rhel fedora"
VERSION_ID="7"
PRETTY_NAME="CentOS Linux 7 (Core)"
ANSI_COLOR="0;31"
CPE_NAME="cpe:/o:centos:centos:7"
HOME_URL="https://www.centos.org/"
BUG_REPORT_URL="https://bugs.centos.org/"

CENTOS_MANTISBT_PROJECT="CentOS-7"
CENTOS_MANTISBT_PROJECT_VERSION="7"
REDHAT_SUPPORT_PRODUCT="centos"
REDHAT_SUPPORT_PRODUCT_VERSION="7"

• Version of Ansible (ansible --version):

ansible 2.7.12
  config file = /VAR/jfmonnier/mvne-cluster/ansible.cfg
  configured module search path = ['/VAR/jfmonnier/mvne-cluster/library']
  ansible python module location = /VAR/jfmonnier/lib/venv/lib/python3.6/site-packages/ansible
  executable location = /VAR/jfmonnier/lib/venv/bin/ansible
  python version = 3.6.6 (default, Aug 13 2018, 18:24:23) [GCC 4.8.5 20150623 (Red Hat 4.8.5-28)]

Kubespray version (commit) (git rev-parse --short HEAD):
7f74906

Network plugin used: Calico

Copy of your inventory file: https://gist.github.com/kesslerdev/d31d11c9c7cdae2ee1d1cf2e9537c651#file-inventory

Command used to invoke ansible:

ansible-playbook -i inventory/mvnecluster/hosts.ini --become --become-user=root -e generic_user=k8s --private-key=mvne-key -e kube_version=v1.14.1 --limit indus5 upgrade-cluster.yml

Output of ansible run:
https://gist.github.com/kesslerdev/d31d11c9c7cdae2ee1d1cf2e9537c651#file-results

Anything else do we need to know:

https://transfer.sh/b5Dmc/logs.tar.gz

[kesslerdev]

after reading the playbook, when upgrading a master, run on first master is mandatory, limit argument is error prone