Remediating nodes with the Poison Pill Operator

You can use the Poison Pill Operator to automatically reboot unhealthy nodes. This remediation strategy minimizes downtime for stateful applications and ReadWriteOnce (RWO) volumes, and restores compute capacity if transient failures occur.

About the Poison Pill Operator

The Poison Pill Operator runs on the cluster nodes and reboots nodes that are identified as unhealthy. The Operator uses the MachineHealthCheck controller to detect the health of a node in the cluster. When a node is identified as unhealthy, the MachineHealthCheck resource creates the PoisonPillRemediation custom resource (CR), which triggers the Poison Pill Operator.

The Poison Pill Operator provides the following capabilities:

  • Minimizes downtime for stateful applications and restores compute capacity if transient failures occur.

  • Independent of any management interface, such as IPMI or an API to provision a node.

Understanding the Poison Pill Operator configuration

The Poison Pill Operator creates the PoisonPillConfig CR with the name poison-pill-config in the Poison Pill Operator’s namespace. You can edit this CR. However, you cannot create a new CR for the Poison Pill Operator.

A change in the PoisonPillConfig CR re-creates the Poison Pill daemon set.

The PoisonPillConfig CR resembles the following YAML file:

  1. apiVersion: poison-pill.medik8s.io/v1alpha1
  2. kind: PoisonPillConfig
  3. metadata:
  4.   name: poison-pill-config
  5.   namespace: openshift-operators
  6. spec:
  7.   safeTimeToAssumeNodeRebootedSeconds: 180 (1)
  8.   watchdogFilePath: /test/watchdog1 (2)
1Specify the timeout duration for the surviving peer, after which the Operator can assume that an unhealthy node has been rebooted. The Operator automatically calculates the lower limit for this value. However, if different nodes have different watchdog timeouts, you must change this value to a higher value.
2Specify the file path of the watchdog device in the nodes. If a watchdog device is unavailable, the PoisonPillConfig CR uses a software reboot.

Installing the Poison Pill Operator by using the web console

You can use the OKD web console to install the Poison Pill Operator.

Prerequisites

  • Log in as a user with cluster-admin privileges.

Procedure

  1. In the OKD web console, navigate to Operators → OperatorHub.

  2. Search for the Poison Pill Operator from the list of available Operators, and then click Install.

  3. Keep the default selection of Installation mode and namespace to ensure that the Operator is installed to the poison-pill namespace.

  4. Click Install.

Verification

To confirm that the installation is successful:

  1. Navigate to the Operators → Installed Operators page.

  2. Check that the Operator is installed in the poison-pill namespace and its status is Succeeded.

If the Operator is not installed successfully:

  1. Navigate to the Operators → Installed Operators page and inspect the Status column for any errors or failures.

  2. Navigate to the Workloads → Pods page and check the logs in any pods in the poison-pill-controller-manager project that are reporting issues.

Installing the Poison Pill Operator by using the CLI

You can use the OpenShift CLI (oc) to install the Poison Pill Operator.

Prerequisites

  • Install the OpenShift CLI (oc).

  • Log in as a user with cluster-admin privileges.

Procedure

  1. Create a Namespace custom resource (CR) for the Poison Pill Operator:

    1. Define the Namespace CR and save the YAML file, for example, poison-pill-namespace.yaml:

      1. apiVersion: v1
      2. kind: Namespace
      3. metadata:
      4.   name: poison-pill

    2. To create the Namespace CR, run the following command:

      1. $ oc create -f poison-pill-namespace.yaml

  2. Create an OperatorGroup CR:

    1. Define the OperatorGroup CR and save the YAML file, for example, poison-pill-operator-group.yaml:

      1. apiVersion: operators.coreos.com/v1
      2. kind: OperatorGroup
      3. metadata:
      4.   name: poison-pill-manager
      5.   namespace: poison-pill
      6. spec:
      7.   targetNamespaces:
      8.   - poison-pill

    2. To create the OperatorGroup CR, run the following command:

      1. $ oc create -f poison-pill-operator-group.yaml

  3. Create a Subscription CR:

    1. Define the Subscription CR and save the YAML file, for example, poison-pill-subscription.yaml:

      1. apiVersion: operators.coreos.com/v1alpha1
      2. kind: Subscription
      3. metadata:
      4.     name: poison-pill-manager
      5.     namespace: poison-pill
      6. spec:
      7.     channel: alpha
      8.     name: poison-pill-manager
      9.     source: redhat-operators
      10.     sourceNamespace: openshift-marketplace
      11.     package: poison-pill-manager

    2. To create the Subscription CR, run the following command:

      1. $ oc create -f poison-pill-subscription.yaml

Verification

  1. Verify that the installation succeeded by inspecting the CSV resource:

    1. $ oc get csv -n poison-pill

    Example output

    1. NAME                   DISPLAY                 VERSION   REPLACES    PHASE
    2. poison-pill.v0.1.4     Poison Pill Operator    0.1.4                 Succeeded

  2. Verify that the Poison Pill Operator is up and running:

    1. $ oc get deploy -n poison-pill

    Example output

    1. NAME                                 READY   UP-TO-DATE   AVAILABLE   AGE
    2. poison-pill-controller-manager       1/1     1            1           10d

  3. Verify that the Poison Pill Operator created the PoisonPillConfig CR:

    1. $ oc get PoisonPillConfig -n poison-pill

    Example output

    1. NAME                 AGE
    2. poison-pill-config   10d

  4. Verify that each poison pill pod is scheduled and running on each worker node:

    1. $ oc get daemonset -n poison-pill

    Example output

    1. NAME             DESIRED   CURRENT   READY   UP-TO-DATE   AVAILABLE   NODE SELECTOR   AGE
    2. poison-pill-ds   2         2         2       2            2           <none>          10d

    This command is unsupported for the control plane nodes.

Configuring machine health checks to use the Poison Pill Operator

Use the following procedure to configure the machine health checks to use the Poison Pill Operator as a remediation provider.

Prerequisites

  • Install the OpenShift CLI (oc).

  • Log in as a user with cluster-admin privileges.

Procedure

  1. Create a PoisonPillRemediationTemplate CR:

    1. Define the PoisonPillRemediationTemplate CR:

      1. apiVersion: poison-pill.medik8s.io/v1alpha1
      2. kind: PoisonPillRemediationTemplate
      3. metadata:
      4.   namespace: openshift-machine-api
      5.   name: poisonpillremediationtemplate-sample
      6. spec:
      7.   template:
      8.     spec: {}

    2. To create the PoisonPillRemediationTemplate CR, run the following command:

      1. $ oc create -f <ppr-name>.yaml

  2. Create or update the MachineHealthCheck CR to point to the PoisonPillRemediationTemplate CR:

    1. Define or update the MachineHealthCheck CR:

      1. apiVersion: machine.openshift.io/v1beta1
      2. kind: MachineHealthCheck
      3. metadata:
      4.   name: machine-health-check
      5.   namespace: openshift-machine-api
      6. spec:
      7.   selector:
      8.     matchLabels:
      9.       machine.openshift.io/cluster-api-machine-role: "worker"
      10.       machine.openshift.io/cluster-api-machine-type: "worker"
      11.   unhealthyConditions:
      12.   - type:    "Ready"
      13.     timeout: "300s"
      14.     status: "False"
      15.   - type:    "Ready"
      16.     timeout: "300s"
      17.     status: "Unknown"
      18.   maxUnhealthy: "40%"
      19.   nodeStartupTimeout: "10m"
      20.   remediationTemplate: (1)
      21.     kind: PoisonPillRemediationTemplate
      22.     apiVersion: poison-pill.medik8s.io/v1alpha1
      23.     name: <poison-pill-remediation-template-sample>
      1Specify the details for the remediation template.

    2. To create a MachineHealthCheck CR, run the following command:

      1. $ oc create -f <file-name>.yaml

    3. To update a MachineHealthCheck CR, run the following command:

      1. $ oc apply -f <file-name>.yaml

Troubleshooting the Poison Pill Operator

General troubleshooting

Issue

You want to troubleshoot issues with the Poison Pill Operator.

Resolution

Check the Operator logs.

Checking the daemon set

Issue

The Poison Pill Operator is installed but the daemon set is not available.

Resolution

Check the Operator logs for errors or warnings.

Unsuccessful remediation

Issue

An unhealthy node was not remediated.

Resolution

Verify that the PoisonPillRemediation CR was created by running the following command:

  1. $ oc get ppr -A

If the MachineHealthCheck controller did not create the PoisonPillRemediation CR when the node turned unhealthy, check the logs of the MachineHealthCheck controller. Additionally, ensure that the MachineHealthCheck CR includes the required specification to use the remediation template.

If the PoisonPillRemediation CR was created, ensure that its name matches the unhealthy node or the machine object.

Daemon set exists even after uninstalling the Poison Pill Operator

Issue

The Poison Pill daemon set exists even after after uninstalling the Operator.

Resolution

To remove the Poison Pill daemon set, manually delete the PoisonPillConfig CR. Run the following command:

  1. $ oc delete ds <poison-pill-daemonset> -n <namespace>

Additional resources

The Poison Pill Operator is supported in a restricted network environment. For more information, see Using Operator Lifecycle Manager on restricted networks.