kubespray/roles/kubernetes/master/tasks/post-upgrade.yml

---
- name: "Post-upgrade | etcd3 upgrade | purge etcd2 k8s data"
  command: "{{ bin_dir }}/etcdctl --endpoints={{ etcd_access_addresses }} rm -r /registry"
  environment:
    ETCDCTL_API: 2
  delegate_to: "{{groups['etcd'][0]}}"
  run_once: true
  when: kube_apiserver_storage_backend == "etcd3" and  needs_etcd_migration|bool|default(false)
Migrate k8s data to etcd3 api store Default backend is now etcd3 (was etcd2). The migration process consists of the following steps: * check if migration is necessary * stop etcd on first etcd server * run migration script * start etcd on first etcd server * stop kube-apiserver until configuration is updated * update kube-apiserver * purge old etcdv2 data 2017-03-06 18:22:13 +08:00			`---`
			`- name: "Post-upgrade \| etcd3 upgrade \| purge etcd2 k8s data"`
			`command: "{{ bin_dir }}/etcdctl --endpoints={{ etcd_access_addresses }} rm -r /registry"`
			`environment:`
			`ETCDCTL_API: 2`
Fix delegate tasks for kubectl and etcdctl 2017-03-23 18:10:42 +08:00			`delegate_to: "{{groups['etcd'][0]}}"`
ensure post-upgrade purge ones only once 2017-03-24 00:19:51 +08:00			`run_once: true`
Migrate k8s data to etcd3 api store Default backend is now etcd3 (was etcd2). The migration process consists of the following steps: * check if migration is necessary * stop etcd on first etcd server * run migration script * start etcd on first etcd server * stop kube-apiserver until configuration is updated * update kube-apiserver * purge old etcdv2 data 2017-03-06 18:22:13 +08:00			`when: kube_apiserver_storage_backend == "etcd3" and needs_etcd_migration\|bool\|default(false)`