From e626cfc8630d27949b33d54c023922d3ed4fc4b1 Mon Sep 17 00:00:00 2001 From: Sebastiaan van Steenis Date: Fri, 26 Oct 2018 23:36:43 +0200 Subject: [PATCH] Add FAQ entry on pod eviction --- content/rancher/v2.x/en/faq/technical/_index.md | 13 +++++++++++++ 1 file changed, 13 insertions(+) diff --git a/content/rancher/v2.x/en/faq/technical/_index.md b/content/rancher/v2.x/en/faq/technical/_index.md index 2ead99f4b86..e82d4ac0ce2 100644 --- a/content/rancher/v2.x/en/faq/technical/_index.md +++ b/content/rancher/v2.x/en/faq/technical/_index.md @@ -123,3 +123,16 @@ When the node is removed from the cluster, and the node is cleaned, you can read ### How can I add additional arguments/binds/environment variables to Kubernetes components in a Rancher Launched Kubernetes cluster? You can add additional arguments/binds/environment variables via the [Config File]({{< baseurl >}}/rancher/v2.x/en/cluster-provisioning/rke-clusters/options/#config-file) option in Cluster Options. For more information, see the [Extra Args, Extra Binds, and Extra Environment Variables]({{< baseurl >}}/rke/v0.1.x/en/config-options/services/services-extras/) in the RKE documentation or browse the [Example Cluster.ymls]({{< baseurl >}}/rke/v0.1.x/en/example-yamls/). + +### Why does it take 5+ minutes for a pod to be rescheduled when a node has failed? + +This is due to a combination of the following default Kubernetes settings: + +* kubelet + * `node-status-update-frequency`: Specifies how often kubelet posts node status to master (default 10s) +* kube-controller-manager + * `node-monitor-period`: The period for syncing NodeStatus in NodeController (default 5s) + * `node-monitor-grace-period`: Amount of time which we allow running Node to be unresponsive before marking it unhealthy (default 40s) + * `pod-eviction-timeout`: The grace period for deleting pods on failed nodes (default 5m0s) + +See [Kubernetes: kubelet](https://kubernetes.io/docs/reference/command-line-tools-reference/kubelet/) and [Kubernetes: kube-controller-manager](https://kubernetes.io/docs/reference/command-line-tools-reference/kube-controller-manager/) for more information on these settings.