From e568b61cd597f1cb714c8e330aca2a2fe6f8d139 Mon Sep 17 00:00:00 2001 From: Qiming Teng Date: Tue, 31 Mar 2020 12:27:24 +0800 Subject: [PATCH] Amend 1.16 kubeadm upgrade guide There is a common issue when upgrading a cluster to 1.16.x from 1.15.y. --- .../kubeadm/kubeadm-upgrade.md | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+) diff --git a/content/en/docs/tasks/administer-cluster/kubeadm/kubeadm-upgrade.md b/content/en/docs/tasks/administer-cluster/kubeadm/kubeadm-upgrade.md index e8b6ae88aa..39eb1b3b23 100644 --- a/content/en/docs/tasks/administer-cluster/kubeadm/kubeadm-upgrade.md +++ b/content/en/docs/tasks/administer-cluster/kubeadm/kubeadm-upgrade.md @@ -237,6 +237,24 @@ The upgrade workflow at high level is the following: This step is not required on additional control plane nodes if the CNI provider runs as a DaemonSet. + One common issue you may encounter if you are using Flannel network is + that after the upgrade, the cluster nodes remain in `NotReady` state. + If you do `kubectl get nodes -o yaml`, you may see following lines in the + output: + + ``` + message:docker: network plugin is not ready: cni config uninitialized + ``` + + In this case, you will need to update the `/etc/cni/net.d/10-flannel.conflist` + file to include the `cniVersion` line, as shown below: + + ``` + "name": "cbr0", + "cniVersion": "0.2.0", + "plugins": [ ... + ``` + 1. Uncordon the control plane node ```shell