Skip to content

Latest commit

 

History

History
88 lines (55 loc) · 5.48 KB

Node not Ready.md

File metadata and controls

88 lines (55 loc) · 5.48 KB
Raw

A Kubernetes node is a physical or virtual machine participating in a Kubernetes cluster, which can e used to run pods. When a node shuts down or crashed, it enters the NotReasy state, meaning it cannot be used to run pods. All stateful pods running on the node then becom unavailable.

Common reasons for a Kubernetes node not ready error include lack of resources on the node, a problem with the kubelet (the agent enabling the Kubernetes control plane to access and control the node), or an error related to kube-proxy(the networking agent on the node).

To identify a Kubernetes node not ready error: run the kubectl get nodes command. Nodes that are not ready will appear like this:

NAME                   STATUS    ROLES   AGE      VERSION
master.example.com     Ready     master  5h       v1.17
node1.example.com      NotReady  compute 5h       v1.17
node2.example.com      Ready     compute 5h       v1.17

We'll provide best practices for diagnosing simple cases of the node not ready error, but more complex cases will require advanced diagnosis and troubleshooting, which is beyond the scope of this article.

Node states

At any given time, a Kubernetes node can be in one of the following states:

  • Ready: able to run pods.
  • NotReady: not operating due to a problem, and cannot run pods.
  • SchedulingDisabled: the node is healthy but has been marked by the cluster as not schedulable.
  • Unknown: If the node controller cannot communicate with the node, it waits a default of 40 seconds, and then sets the node status to unknown.

If a note is in the NodeReady state, it indicates that the kubelet is installed on the node, but Kubernetes has detected a problem on the node that prevents it from running pods.

Troubleshooing Node Not Ready Error

Here are some common reasons that a Kubernetes node may enter the NotRead state:

Lack of System Resources

A node must have enough dist space, memory, and processing power to run Kubernetes workloads.

If non-Kubernetes processes on the node are taking up too many resources, or if there are too many processes running on the node, it can be marked by the control plane as NotReady

Run kubectl describe node and look in the Conditions section to see if resources are missing on the node:

  • MemoryPressure: node is running out of memory.
  • DiskPressure: node is running out of disk space.
  • PIDPressure: node is running too many processes.

Here are a few ways to resolve a system resource issue on the node:

  • Identify which non-Kubernetes processes are running on the node. If there are any, shut them down or reduce them to a minimum to conserve resources.
  • Run a malware scan—there may be hidden malicious processes taking up system resources.
  • Upgrade the node.
  • Check for hardware issues or misconfigurations and resolve them.

Kubelet Issue

The Kubelet must run on each node to enable it to perticipate in the cluster. If the kubelet crashes or stops on a node, it cannot communicate with the API server and the node goes into a not ready state.

Run kubectl describe node [name] and look in the Conditions section. If all the conditions are unknown, this indicates the kubelet is down.

To resolve a kubelet issue, SSH into the node and run the command systemctl status kubelet

Look at the value of the Active field:

  • active (running) means the kubelet is actually operational, look for the problem elsewhere.
  • active (exited) means the kubelet was exited, probably in error. Restart it.>
  • inactive (dead) means the kubelet crashed. To identify why, run the command journalctl -u kubelet and examine the kubelet logs.

Kube-proxy Issue

kube-proxy runs on every node and is responsible for regulating network traffic between the node and other entities inside and outside the cluster. If kube-proxy stops running for any reason, the node goes into a not ready state.

Run kubectl get pods - kube-system to show pods belonging to the Kubernetes system.

Try looking in the following places to identify what is the issue with kube-proxy:

  • Run the command kubectl describe pod using the name of the kube-proxy pod that failed, and check the Events section in the output.
  • Run the command kubectl logs [pod-name] -n kube-system to see a full log of the failing kube-proxy pod.
  • Run the command kubectl describe daemonset kube-proxy -n kube-system to see the status of the kube-proxy daemonset, which is responsible for ensuring there is a kube-proxy running on every Kubernetes node.

Please note that these procedures can help you gather more information about the problem, but additional steps may be needed to resolve the problem. If one of the quick fixes above did not work, you’ll need to undertake a more complex, non-linear diagnosis procedure to identify which parts of the Kubernetes environment contribute to the node not ready problem and resolve it.

Connectivity Issue

Even if a node is configured perfectly, but it has no network connectivity, Kubernetes treats the node as not ready. This could be due to a disconnected network cable, no Internet access, or misconfigured networking on the machine.

Run kubectl describe node [name] and look in the Conditions section. if the NetworkUnavailable flag is true, this means the node has a connectivity issue.

reference