×

Understanding how to evacuate pods on nodes

Evacuating pods allows you to migrate all or selected pods from a given node or nodes.

You can only evacuate pods backed by a replication controller. The replication controller creates new pods on other nodes and removes the existing pods from the specified node(s).

Bare pods, meaning those not backed by a replication controller, are unaffected by default. You can evacuate a subset of pods by specifying a pod-selector. Pod selectors are based on labels, so all the pods with the specified label will be evacuated.

Procedure
  1. Mark the nodes unschedulable before performing the pod evacuation.

    1. Mark the node as unschedulable:

      $ oc adm cordon <node1>
      Example output
      node/<node1> cordoned
    2. Check that the node status is Ready,SchedulingDisabled:

      $ oc get node <node1>
      Example output
      NAME        STATUS                     ROLES     AGE       VERSION
      <node1>     Ready,SchedulingDisabled   worker    1d        v1.24.0
  2. Evacuate the pods using one of the following methods:

    • Evacuate all or selected pods on one or more nodes:

      $ oc adm drain <node1> <node2> [--pod-selector=<pod_selector>]
    • Force the deletion of bare pods using the --force option. When set to true, deletion continues even if there are pods not managed by a replication controller, replica set, job, daemon set, or stateful set:

      $ oc adm drain <node1> <node2> --force=true
    • Set a period of time in seconds for each pod to terminate gracefully, use --grace-period. If negative, the default value specified in the pod will be used:

      $ oc adm drain <node1> <node2> --grace-period=-1
    • Ignore pods managed by daemon sets using the --ignore-daemonsets flag set to true:

      $ oc adm drain <node1> <node2> --ignore-daemonsets=true
    • Set the length of time to wait before giving up using the --timeout flag. A value of 0 sets an infinite length of time:

      $ oc adm drain <node1> <node2> --timeout=5s
    • Delete pods even if there are pods using emptyDir volumes by setting the --delete-emptydir-data flag to true. Local data is deleted when the node is drained:

      $ oc adm drain <node1> <node2> --delete-emptydir-data=true
    • List objects that will be migrated without actually performing the evacuation, using the --dry-run option set to true:

      $ oc adm drain <node1> <node2>  --dry-run=true

      Instead of specifying specific node names (for example, <node1> <node2>), you can use the --selector=<node_selector> option to evacuate pods on selected nodes.

  3. Mark the node as schedulable when done.

    $ oc adm uncordon <node1>

Understanding how to update labels on nodes

You can update any label on a node.

Node labels are not persisted after a node is deleted even if the node is backed up by a Machine.

Any change to a MachineSet object is not applied to existing machines owned by the machine set. For example, labels edited or added to an existing MachineSet object are not propagated to existing machines and nodes associated with the machine set.

  • The following command adds or updates labels on a node:

    $ oc label node <node> <key_1>=<value_1> ... <key_n>=<value_n>

    For example:

    $ oc label nodes webconsole-7f7f6 unhealthy=true
  • The following command updates all pods in the namespace:

    $ oc label pods --all <key_1>=<value_1>

    For example:

    $ oc label pods --all status=unhealthy

Understanding how to mark nodes as unschedulable or schedulable

By default, healthy nodes with a Ready status are marked as schedulable, meaning that new pods are allowed for placement on the node. Manually marking a node as unschedulable blocks any new pods from being scheduled on the node. Existing pods on the node are not affected.

  • The following command marks a node or nodes as unschedulable:

    Example output
    $ oc adm cordon <node>

    For example:

    $ oc adm cordon node1.example.com
    Example output
    node/node1.example.com cordoned
    
    NAME                 LABELS                                        STATUS
    node1.example.com    kubernetes.io/hostname=node1.example.com      Ready,SchedulingDisabled
  • The following command marks a currently unschedulable node or nodes as schedulable:

    $ oc adm uncordon <node1>

    Alternatively, instead of specifying specific node names (for example, <node>), you can use the --selector=<node_selector> option to mark selected nodes as schedulable or unschedulable.

Configuring control plane nodes as schedulable

You can configure control plane nodes (also known as the master nodes) to be schedulable, meaning that new pods are allowed for placement on the master nodes. By default, control plane nodes are not schedulable.

You can set the masters to be schedulable, but must retain the worker nodes.

You can deploy OKD with no worker nodes on a bare metal cluster. In this case, the control plane nodes are marked schedulable by default.

You can allow or disallow control plane nodes to be schedulable by configuring the mastersSchedulable field.

+

When you configure control plane nodes from the default unschedulable to schedulable, additional subscriptions are required. This is because control plane nodes then become worker nodes.

Procedure
  1. Edit the schedulers.config.openshift.io resource.

    $ oc edit schedulers.config.openshift.io cluster
  2. Configure the mastersSchedulable field.

    apiVersion: config.openshift.io/v1
    kind: Scheduler
    metadata:
      creationTimestamp: "2019-09-10T03:04:05Z"
      generation: 1
      name: cluster
      resourceVersion: "433"
      selfLink: /apis/config.openshift.io/v1/schedulers/cluster
      uid: a636d30a-d377-11e9-88d4-0a60097bee62
    spec:
      mastersSchedulable: false (1)
      policy:
        name: ""
    status: {}
    1 Set to true to allow control plane nodes to be schedulable, or false to disallow control plane nodes to be schedulable.
  3. Save the file to apply the changes.

Deleting nodes

Deleting nodes from a cluster

When you delete a node using the CLI, the node object is deleted in Kubernetes, but the pods that exist on the node are not deleted. Any bare pods not backed by a replication controller become inaccessible to OKD. Pods backed by replication controllers are rescheduled to other available nodes. You must delete local manifest pods.

Procedure

To delete a node from the OKD cluster, edit the appropriate MachineSet object:

If you are running cluster on bare metal, you cannot delete a node by editing MachineSet objects. Machine sets are only available when a cluster is integrated with a cloud provider. Instead you must unschedule and drain the node before manually deleting it.

  1. View the machine sets that are in the cluster:

    $ oc get machinesets -n openshift-machine-api

    The machine sets are listed in the form of <clusterid>-worker-<aws-region-az>.

  2. Scale the machine set:

    $ oc scale --replicas=2 machineset <machineset> -n openshift-machine-api

For more information on scaling your cluster using a machine set, see Manually scaling a machine set.

Deleting nodes from a bare metal cluster

When you delete a node using the CLI, the node object is deleted in Kubernetes, but the pods that exist on the node are not deleted. Any bare pods not backed by a replication controller become inaccessible to OKD. Pods backed by replication controllers are rescheduled to other available nodes. You must delete local manifest pods.

Procedure

Delete a node from an OKD cluster running on bare metal by completing the following steps:

  1. Mark the node as unschedulable:

    $ oc adm cordon <node_name>
  2. Drain all pods on the node:

    $ oc adm drain <node_name> --force=true

    This step might fail if the node is offline or unresponsive. Even if the node does not respond, it might still be running a workload that writes to shared storage. To avoid data corruption, power down the physical hardware before you proceed.

  3. Delete the node from the cluster:

    $ oc delete node <node_name>

    Although the node object is now deleted from the cluster, it can still rejoin the cluster after reboot or if the kubelet service is restarted. To permanently delete the node and all its data, you must decommission the node.

  4. If you powered down the physical hardware, turn it back on so that the node can rejoin the cluster.

Setting SELinux booleans

OKD allows you to enable and disable an SELinux boolean on a Fedora CoreOS (FCOS) node. The following procedure explains how to modify SELinux booleans on nodes using the Machine Config Operator (MCO). This procedure uses container_manage_cgroup as the example boolean. You can modify this value to whichever boolean you need.

Prerequisites
  • You have installed the OpenShift CLI (oc).

Procedure
  1. Create a new YAML file with a MachineConfig object, displayed in the following example:

    apiVersion: machineconfiguration.openshift.io/v1
    kind: MachineConfig
    metadata:
      labels:
        machineconfiguration.openshift.io/role: worker
      name: 99-worker-setsebool
    spec:
      config:
        ignition:
          version: 2.2.0
        systemd:
          units:
          - contents: |
              [Unit]
              Description=Set SELinux booleans
              Before=kubelet.service
    
              [Service]
              Type=oneshot
              ExecStart=/sbin/setsebool container_manage_cgroup=on
              RemainAfterExit=true
    
              [Install]
              WantedBy=multi-user.target graphical.target
            enabled: true
            name: setsebool.service
  2. Create the new MachineConfig object by running the following command:

    $ oc create -f 99-worker-setsebool.yaml

Applying any changes to the MachineConfig object causes all affected nodes to gracefully reboot after the change is applied.

Adding kernel arguments to nodes

In some special cases, you might want to add kernel arguments to a set of nodes in your cluster. This should only be done with caution and clear understanding of the implications of the arguments you set.

Improper use of kernel arguments can result in your systems becoming unbootable.

Examples of kernel arguments you could set include:

  • enforcing=0: Configures Security Enhanced Linux (SELinux) to run in permissive mode. In permissive mode, the system acts as if SELinux is enforcing the loaded security policy, including labeling objects and emitting access denial entries in the logs, but it does not actually deny any operations. While not supported for production systems, permissive mode can be helpful for debugging.

  • nosmt: Disables symmetric multithreading (SMT) in the kernel. Multithreading allows multiple logical threads for each CPU. You could consider nosmt in multi-tenant environments to reduce risks from potential cross-thread attacks. By disabling SMT, you essentially choose security over performance.

See Kernel.org kernel parameters for a list and descriptions of kernel arguments.

In the following procedure, you create a MachineConfig object that identifies:

  • A set of machines to which you want to add the kernel argument. In this case, machines with a worker role.

  • Kernel arguments that are appended to the end of the existing kernel arguments.

  • A label that indicates where in the list of machine configs the change is applied.

Prerequisites
  • Have administrative privilege to a working OKD cluster.

Procedure
  1. List existing MachineConfig objects for your OKD cluster to determine how to label your machine config:

    $ oc get MachineConfig
    Example output
    NAME                                               GENERATEDBYCONTROLLER                      IGNITIONVERSION   AGE
    00-master                                          52dd3ba6a9a527fc3ab42afac8d12b693534c8c9   3.2.0             33m
    00-worker                                          52dd3ba6a9a527fc3ab42afac8d12b693534c8c9   3.2.0             33m
    01-master-container-runtime                        52dd3ba6a9a527fc3ab42afac8d12b693534c8c9   3.2.0             33m
    01-master-kubelet                                  52dd3ba6a9a527fc3ab42afac8d12b693534c8c9   3.2.0             33m
    01-worker-container-runtime                        52dd3ba6a9a527fc3ab42afac8d12b693534c8c9   3.2.0             33m
    01-worker-kubelet                                  52dd3ba6a9a527fc3ab42afac8d12b693534c8c9   3.2.0             33m
    99-master-generated-registries                     52dd3ba6a9a527fc3ab42afac8d12b693534c8c9   3.2.0             33m
    99-master-ssh                                                                                 3.2.0             40m
    99-worker-generated-registries                     52dd3ba6a9a527fc3ab42afac8d12b693534c8c9   3.2.0             33m
    99-worker-ssh                                                                                 3.2.0             40m
    rendered-master-23e785de7587df95a4b517e0647e5ab7   52dd3ba6a9a527fc3ab42afac8d12b693534c8c9   3.2.0             33m
    rendered-worker-5d596d9293ca3ea80c896a1191735bb1   52dd3ba6a9a527fc3ab42afac8d12b693534c8c9   3.2.0             33m
  2. Create a MachineConfig object file that identifies the kernel argument (for example, 05-worker-kernelarg-selinuxpermissive.yaml)

    apiVersion: machineconfiguration.openshift.io/v1
    kind: MachineConfig
    metadata:
      labels:
        machineconfiguration.openshift.io/role: worker(1)
      name: 05-worker-kernelarg-selinuxpermissive(2)
    spec:
      config:
        ignition:
          version: 3.2.0
      kernelArguments:
        - enforcing=0(3)
    1 Applies the new kernel argument only to worker nodes.
    2 Named to identify where it fits among the machine configs (05) and what it does (adds a kernel argument to configure SELinux permissive mode).
    3 Identifies the exact kernel argument as enforcing=0.
  3. Create the new machine config:

    $ oc create -f 05-worker-kernelarg-selinuxpermissive.yaml
  4. Check the machine configs to see that the new one was added:

    $ oc get MachineConfig
    Example output
    NAME                                               GENERATEDBYCONTROLLER                      IGNITIONVERSION   AGE
    00-master                                          52dd3ba6a9a527fc3ab42afac8d12b693534c8c9   3.2.0             33m
    00-worker                                          52dd3ba6a9a527fc3ab42afac8d12b693534c8c9   3.2.0             33m
    01-master-container-runtime                        52dd3ba6a9a527fc3ab42afac8d12b693534c8c9   3.2.0             33m
    01-master-kubelet                                  52dd3ba6a9a527fc3ab42afac8d12b693534c8c9   3.2.0             33m
    01-worker-container-runtime                        52dd3ba6a9a527fc3ab42afac8d12b693534c8c9   3.2.0             33m
    01-worker-kubelet                                  52dd3ba6a9a527fc3ab42afac8d12b693534c8c9   3.2.0             33m
    05-worker-kernelarg-selinuxpermissive                                                         3.2.0             105s
    99-master-generated-registries                     52dd3ba6a9a527fc3ab42afac8d12b693534c8c9   3.2.0             33m
    99-master-ssh                                                                                 3.2.0             40m
    99-worker-generated-registries                     52dd3ba6a9a527fc3ab42afac8d12b693534c8c9   3.2.0             33m
    99-worker-ssh                                                                                 3.2.0             40m
    rendered-master-23e785de7587df95a4b517e0647e5ab7   52dd3ba6a9a527fc3ab42afac8d12b693534c8c9   3.2.0             33m
    rendered-worker-5d596d9293ca3ea80c896a1191735bb1   52dd3ba6a9a527fc3ab42afac8d12b693534c8c9   3.2.0             33m
  5. Check the nodes:

    $ oc get nodes
    Example output
    NAME                           STATUS                     ROLES    AGE   VERSION
    ip-10-0-136-161.ec2.internal   Ready                      worker   28m   v1.20.0
    ip-10-0-136-243.ec2.internal   Ready                      master   34m   v1.20.0
    ip-10-0-141-105.ec2.internal   Ready,SchedulingDisabled   worker   28m   v1.20.0
    ip-10-0-142-249.ec2.internal   Ready                      master   34m   v1.20.0
    ip-10-0-153-11.ec2.internal    Ready                      worker   28m   v1.20.0
    ip-10-0-153-150.ec2.internal   Ready                      master   34m   v1.20.0

    You can see that scheduling on each worker node is disabled as the change is being applied.

  6. Check that the kernel argument worked by going to one of the worker nodes and listing the kernel command line arguments (in /proc/cmdline on the host):

    $ oc debug node/ip-10-0-141-105.ec2.internal
    Example output
    Starting pod/ip-10-0-141-105ec2internal-debug ...
    To use host binaries, run `chroot /host`
    
    sh-4.2# cat /host/proc/cmdline
    BOOT_IMAGE=/ostree/rhcos-... console=tty0 console=ttyS0,115200n8
    rootflags=defaults,prjquota rw root=UUID=fd0... ostree=/ostree/boot.0/rhcos/16...
    coreos.oem.id=qemu coreos.oem.id=ec2 ignition.platform.id=ec2 enforcing=0
    
    sh-4.2# exit

    You should see the enforcing=0 argument added to the other kernel arguments.

Additional resources