System Tools is a tool to perform operational tasks on Rancher Launched Kubernetes clusters or installations of Rancher on an RKE cluster. The tasks include:

  • Collect logging and system metrics from nodes.
  • Remove Kubernetes resources created by Rancher.

The following commands are available:

Command Description
logs Collect Kubernetes cluster component logs from nodes.
stats Stream system metrics from nodes.
remove Remove Kubernetes resources created by Rancher.

Download System Tools

You can download the latest version of System Tools from the GitHub releases page. Download the version of system-tools for the OS that you are using to interact with the cluster.

Operating System Filename
MacOS system-tools_darwin-amd64
Linux system-tools_linux-amd64
Windows system-tools_windows-amd64.exe

After you download the tools, complete the following actions:

  1. Rename the file to system-tools.

  2. Give the file executable permissions by running the following command:

    Using Windows? The file is already an executable, you can skip this step.

    chmod +x system-tools

Logs

The logs subcommand will collect log files of core Kubernetes cluster components from nodes in Rancher-launched Kubernetes clusters or nodes on an RKE Kubernetes cluster that Rancher is installed on.. See Troubleshooting for a list of core Kubernetes cluster components.

System Tools will use the provided kubeconfig file to deploy a DaemonSet, that will copy all the logfiles from the core Kubernetes cluster components and add them to a single tar file (cluster-logs.tar by default). If you only want to collect logging from a single node, you can specify the node by using --node NODENAME or -n NODENAME.

Usage

./system-tools_darwin-amd64 logs --kubeconfig <KUBECONFIG>

The following are the options for the logs command:

Option Description
--kubeconfig <KUBECONFIG_PATH>, -c <KUBECONFIG_PATH> The cluster’s kubeconfig file.
--output <FILENAME>, -o cluster-logs.tar Name of the created tarball containing the logs. If no output filename is defined, the options defaults to cluster-logs.tar.
--node <NODENAME>, -n node1 Specify the nodes to collect the logs from. If no node is specified, logs from all nodes in the cluster will be collected.

Stats

The stats subcommand will display system metrics from nodes in Rancher-launched Kubernetes clusters or nodes in an RKE Kubernetes cluster that Rancher is installed on..

System Tools will deploy a DaemonSet, and run a predefined command based on sar (System Activity Report) to show system metrics.

Usage

./system-tools_darwin-amd64 stats --kubeconfig <KUBECONFIG>

The following are the options for the stats command:

Option Description
--kubeconfig <KUBECONFIG_PATH>, -c <KUBECONFIG_PATH> The cluster’s kubeconfig file.
--node <NODENAME>, -n node1 Specify the nodes to display the system metrics from. If no node is specified, logs from all nodes in the cluster will be displayed.
--stats-command value, -s value The command to run to display the system metrics. If no command is defined, the options defaults to /usr/bin/sar -u -r -F 1 1.

Remove

Warning: This command will remove data from your etcd nodes. Make sure you have created a backup of etcd before executing the command.

When you install Rancher on a Kubernetes cluster, it will create Kubernetes resources to run and to store configuration data. If you want to remove Rancher from your cluster, you can use the remove subcommand to remove the Kubernetes resources. When you use the remove subcommand, the following resources will be removed:

  • The Rancher deployment namespace (cattle-system by default).
  • Any serviceAccount, clusterRoles, and clusterRoleBindings that Rancher applied the cattle.io/creator:norman label to. Rancher applies this label to any resource that it creates as of v2.1.0.
  • Labels, annotations, and finalizers.
  • Rancher Deployment.
  • Machines, clusters, projects, and user custom resource deployments (CRDs).
  • All resources create under the management.cattle.io API Group.
  • All CRDs created by Rancher v2.x.

Using 2.0.8 or Earlier?

These versions of Rancher do not automatically delete the serviceAccount, clusterRole, and clusterRoleBindings resources after the job runs. You’ll have to delete them yourself.

Usage

When you run the command below, all the resources listed above will be removed from the cluster.

Warning: This command will remove data from your etcd nodes. Make sure you have created a backup of etcd before executing the command.

./system-tools remove --kubeconfig <KUBECONFIG> --namespace <NAMESPACE>

The following are the options for the remove command:

Option Description
--kubeconfig <KUBECONFIG_PATH>, -c <KUBECONFIG_PATH> The cluster’s kubeconfig file
--namespace <NAMESPACE>, -n cattle-system Rancher 2.x deployment namespace (<NAMESPACE>). If no namespace is defined, the options defaults to cattle-system.
--force Skips the interactive removal confirmation and removes the Rancher deployment without prompt.