Devtron K8s Dashboard
  • Getting Started
    • Overview of Dashboard
    • Prerequisites
    • Install Modern Kubernetes Dashboard
  • User Guide
    • Explore Kubernetes Resources
      • Overview Page
      • Discover and Manage Resources
      • Nodes and Operations
      • Pod Management and Debugging
      • Cluster Terminal
      • Add Monitoring Dashboards/Graphs
      • Run Kubectl Commands Locally
    • Use Resource Watcher
    • Manage Helm Apps
    • Manage Argo CD Apps
    • Manage Flux CD Apps
    • Chart Store
      • Examples
        • Deploying MySQL Helm Chart
        • Deploying MongoDB Helm Chart
  • Operator Guide
    • Projects
    • Clusters
    • OCI Registry
    • Chart Repositories
    • Manage Authorization (RBAC)
      • SSO Login Services
        • Google
        • GitHub
        • GitLab
        • Microsoft
        • LDAP
        • OIDC
          • Keycloak
          • Okta
        • OpenShift
      • User Permissions
      • Permission Groups
      • API Tokens
    • External Links
    • Catalog Framework
    • Charts and Chart Store
    • Show/Hide Argo CD App Listing
    • Show/Hide Flux CD App Listing
    • Configure GUI Schema for Manifests
    • Configure Lock Schema for Manifests
  • Resources
    • Glossary
    • FAQ
Powered by GitBook
On this page
  • Debug a Node
  • Cordon a Node
  • Drain a Node
  • Taint a Node
  • Edit a Node Config
  • Delete a Node
Export as PDF
  1. User Guide
  2. Explore Kubernetes Resources

Nodes and Operations

PreviousDiscover and Manage ResourcesNextPod Management and Debugging

Last updated 6 months ago

You can see the list of nodes available in your cluster. Typically you have several nodes in a cluster; in a learning or resource-limited environment, you might have only one node.

The components on a typical node include the kubelet, a container runtime, and the kube-proxy.

If you have multiple nodes, you can search a node by name or label in the search bar. The search result will display the following information about the node. To display a parameter of a node, use Columns on the right side, select the parameter to display from the drop-down list, and click Apply.

Figure 1: Searching and Filtering Nodes
Fields
Description

Node

Alphanumeric name of the node

Status

Status of a node. It can be either Ready or Not Ready.

Roles

Shows the roles of a node, e.g., agent

Errors

Shows the number of errors in nodes (if any)

K8s Version

Shows the version of Kubernetes cluster

Node Group

Shows which collection of worker nodes it belongs to

No. of Pods

Shows the total number of pods present in the node

Clicking on a node shows you a number of details such as:

  • CPU Usage and Memory Usage of Node

  • CPU Usage and Memory Usage of Each Pod

  • Number of Pods in the Node

  • List of Pods

  • Age of Pods

  • Labels, Annotations, and Taints

  • Node IP

Further using the Devtron UI, you will be able to:

Why Are Node Operations Required?

Your applications run on pods, and pods run on nodes. But sometimes, Kubernetes scheduler cannot deploy a pod on a node for several reasons, e.g., node is not ready, node is not reachable, network is unavailable, etc. In such cases, node operations help you manage the nodes better.

Who Can Perform These Actions?

Users need to have super-admin permission to perform node operations.

Debug a Node

  • Click Debug.

  • Debug a node by selecting the terminal shell, i.e., bash or sh.


Cordon a Node

  • Click Cordon.

  • A confirmation dialog box will appear, click Cordon Node to proceed.

The status of the node shows SchedulingDisabled with Unschedulable parameter set as true.

Similarly, you can uncordon a node by clicking Uncordon. After a node is uncordoned, new pods can be scheduled on the node.


Drain a Node

After the node is drained, all pods (including those managed by DaemonSets) in the node will be automatically drained to other nodes in the cluster, and the drained node will be set to cordoned status.

  • Click Drain.

  • A confirmation dialog box will appear, click Drain Node to proceed.

You can also select from the following conditions before draining a node:

Name
Usage

Grace Period

Period of time in seconds given to each pod to terminate gracefully. If negative, the default value specified in the pod will be used.

Delete empty directory data

Enabling this field will delete the pods using empty directory data when the node is drained.

Disable eviction (use with caution)

Enabling this field will force drain to use delete, even if eviction is supported. This will bypass checking PodDisruptionBudgets. Note: Make sure to use with caution.

Force drain

Enabling this field will force drain a node even if there are pods that do not declare a controller.

Ignore DaemonSets

Enabling this field will ignore DaemonSet-managed pods.


Taint a Node

Taints are key:value pairs associated with effect. After you add taints to nodes, you can set tolerations on a pod to allow the pod to be scheduled to nodes with certain taints. When you taint a node, it will repel all the pods except those that have a toleration for that taint. A node can have one or many taints associated with it.

Note: Make sure to check taint validations before you add a taint.

  • Click Edit taints.

  • Click Save.

You can also add more taints using + Add taint button, or delete the existing taint by using the delete icon.

Additional Reference


Edit a Node Config

This allows you to directly edit any node. It will open the editor which contains all the configuration settings in which the default format is YAML. You can edit multiple objects, although changes are applied one at a time.

  • Go to the YAML tab and click Edit YAML.

  • Make the changes using the editor.

  • Click Review & Save changes to compare the changes in the YAML file.

  • Click Apply changes to confirm.


Delete a Node

You can also delete a node by clicking the Delete button present on the right-hand side.

The node will be deleted from the cluster.

Figure 2: Checking Node Summary

You can debug a node via by selecting your namespace and image from the list that has all CLI utilities like kubectl, helm, netshoot etc. or can use a custom image, which is publicly available.

Figure 3a: Debugging a Node
Figure 3b: Debug Terminal

Cordoning a node means making the node unschedulable. After , new pods cannot be scheduled on this node.

Figure 4a: Visual Representation - Cordoning a Node
Figure 4b: Cordoning a Node
Figure 4c: Cordon Confirmation

Before performing maintenance on a node, evicts all of your pods safely from a node. Safe evictions allow the pod’s containers to gracefully terminate and honour the PodDisruptionBudgets you have specified (if relevant).

Figure 5a: Visual Representation - Draining a Node
Figure 5b: Draining a Node
Figure 5c: Drain Confirmation
Figure 6a: Visual Representation - Tainting a Node
Figure 6b: Tainting a Node

Enter the key:value pairs and select the from the drop-down list.

Figure 6c: Adding Taints

to read about taint effects.

Figure 7: Editing Node Config
Figure 8a: Deleting a Node
Figure 8b: Delete Confirmation

You can also access from your node.

Cluster Terminal
cordoning a node
draining a node
Click here
Cluster Terminal
Debug a Node
Cordon a Node
Drain a Node
Taint a Node
Edit a Node Config
Delete a Node
taint effect