×

Understanding low latency

The emergence of Edge computing in the area of Telco / 5G plays a key role in reducing latency and congestion problems and improving application performance.

Simply put, latency determines how fast data (packets) moves from the sender to receiver and returns to the sender after processing by the receiver. Obviously, maintaining a network architecture with the lowest possible delay of latency speeds is key for meeting the network performance requirements of 5G. Compared to 4G technology, with an average latency of 50ms, 5G is targeted to reach latency numbers of 1ms or less. This reduction in latency boosts wireless throughput by a factor of 10.

Many of the deployed applications in the Telco space require low latency that can only tolerate zero packet loss. Tuning for zero packet loss helps mitigate the inherent issues that degrade network performance. For more information, see Tuning for Zero Packet Loss in Red Hat OpenStack Platform (RHOSP).

The Edge computing initiative also comes in to play for reducing latency rates. Think of it as literally being on the edge of the cloud and closer to the user. This greatly reduces the distance between the user and distant data centers, resulting in reduced application response times and performance latency.

Administrators must be able to manage their many Edge sites and local services in a centralized way so that all of the deployments can run at the lowest possible management cost. They also need an easy way to deploy and configure certain nodes of their cluster for real-time low latency and high-performance purposes. Low latency nodes are useful for applications such as Cloud-native Network Functions (CNF) and Data Plane Development Kit (DPDK).

OpenShift Container Platform currently provides mechanisms to tune software on an OpenShift Container Platform cluster for real-time running and low latency (around <20 microseconds reaction time). This includes tuning the kernel and OpenShift Container Platform set values, installing a kernel, and reconfiguring the machine. But this method requires setting up four different Operators and performing many configurations that, when done manually, is complex and could be prone to mistakes.

OpenShift Container Platform provides a Performance Addon Operator to implement automatic tuning in order to achieve low latency performance for OpenShift applications. The cluster administrator uses this performance profile configuration that makes it easier to make these changes in a more reliable way. The administrator can specify whether to update the kernel to kernel-rt, the CPUs that will be reserved for housekeeping, and the CPUs that will be used for running the workloads.

Installing the Performance Addon Operator

Performance Addon Operator provides the ability to enable advanced node performance tunings on a set of nodes. As a cluster administrator, you can install Performance Addon Operator using the OpenShift Container Platform CLI or the web console.

Installing the Operator using the CLI

As a cluster administrator, you can install the Operator using the CLI.

Prerequisites
  • A cluster installed on bare-metal hardware.

  • Install the OpenShift CLI (oc).

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

Procedure
  1. Create a namespace for the Performance Addon Operator by completing the following actions:

    1. Create the following Namespace Custom Resource (CR) that defines the openshift-performance-addon-operator namespace, and then save the YAML in the pao-namespace.yaml file:

      apiVersion: v1
      kind: Namespace
      metadata:
        name: openshift-performance-addon-operator
    2. Create the namespace by running the following command:

      $ oc create -f pao-namespace.yaml
  2. Install the Performance Addon Operator in the namespace you created in the previous step by creating the following objects:

    1. Create the following OperatorGroup CR and save the YAML in the pao-operatorgroup.yaml file:

      apiVersion: operators.coreos.com/v1
      kind: OperatorGroup
      metadata:
        name: openshift-performance-addon-operator
        namespace: openshift-performance-addon-operator
      spec:
        targetNamespaces:
        - openshift-performance-addon-operator
    2. Create the OperatorGroup CR by running the following command:

      $ oc create -f pao-operatorgroup.yaml
    3. Run the following command to get the channel value required for the next step.

      $ oc get packagemanifest performance-addon-operator -n openshift-marketplace -o jsonpath='{.status.defaultChannel}'
      Example output
      4.6
    4. Create the following Subscription CR and save the YAML in the pao-sub.yaml file:

      Example Subscription
      apiVersion: operators.coreos.com/v1alpha1
      kind: Subscription
      metadata:
        name: openshift-performance-addon-operator-subscription
        namespace: openshift-performance-addon-operator
      spec:
        channel: "<channel>" (1)
        name: performance-addon-operator
        source: redhat-operators (2)
        sourceNamespace: openshift-marketplace
      1 Specify the value from you obtained in the previous step for the .status.defaultChannel parameter.
      2 You must specify the redhat-operators value.
    5. Create the Subscription object by running the following command:

      $ oc create -f pao-sub.yaml
    6. Change to the openshift-performance-addon-operator project:

      $ oc project openshift-performance-addon-operator

Installing the Performance Addon Operator using the web console

As a cluster administrator, you can install the Performance Addon Operator using the web console.

You must create the Namespace CR and OperatorGroup CR as mentioned in the previous section.

Procedure
  1. Install the Performance Addon Operator using the OpenShift Container Platform web console:

    1. In the OpenShift Container Platform web console, click OperatorsOperatorHub.

    2. Choose Performance Addon Operator from the list of available Operators, and then click Install.

    3. On the Install Operator page, under A specific namespace on the cluster select openshift-performance-addon-operator. Then, click Install.

  2. Optional: Verify that the performance-addon-operator installed successfully:

    1. Switch to the OperatorsInstalled Operators page.

    2. Ensure that Performance Addon Operator is listed in the openshift-operators project with a Status of Succeeded.

      During installation an Operator might display a Failed status. If the installation later succeeds with a Succeeded message, you can ignore the Failed message.

      If the Operator does not appear as installed, you can troubleshoot further:

      • Go to the OperatorsInstalled Operators page and inspect the Operator Subscriptions and Install Plans tabs for any failure or errors under Status.

      • Go to the WorkloadsPods page and check the logs for pods in the openshift-operators project.

Upgrading Performance Addon Operator

You can manually upgrade to the next minor version of Performance Addon Operator and monitor the status of an update by using the web console.

About upgrading Performance Addon Operator

  • You can upgrade to the next minor version of Performance Addon Operator by using the OpenShift web console to change the channel of your Operator subscription.

  • You can enable automatic z-stream updates during Performance Addon Operator installation.

  • Updates are delivered via the Marketplace Operator, which is deployed during OpenShift Container Platform installation.The Marketplace Operator makes external Operators available to your cluster.

  • The amount of time an update takes to complete depends on your network connection. Most automatic updates complete within fifteen minutes.

How Performance Addon Operator upgrades affect your cluster

  • Neither the low latency tuning nor huge pages are affected.

  • Updating the Operator should not cause any unexpected reboots.

Upgrading Performance Addon Operator to the next minor version

You can manually upgrade Performance Addon Operator to the next minor version by using the OpenShift Container Platform web console to change the channel of your Operator subscription.

Prerequisites
  • Access to the cluster as a user with the cluster-admin role.

Procedure
  1. Access the OpenShift web console and navigate to Operators → Installed Operators.

  2. Click Performance Addon Operator to open the Operator Details page.

  3. Click the Subscription tab to open the Subscription Overview page.

  4. In the Channel pane, click the pencil icon on the right side of the version number to open the Change Subscription Update Channel window.

  5. Select the next minor version. For example, if you want to upgrade to Performance Addon Operator 4.6, select 4.6.

  6. Click Save.

  7. Check the status of the upgrade by navigating to Operators → Installed Operators. You can also check the status by running the following oc command:

    $ oc get csv -n openshift-performance-addon-operator

Monitoring upgrade status

The best way to monitor Performance Addon Operator upgrade status is to watch the ClusterServiceVersion (CSV) PHASE. You can also monitor the CSV conditions in the web console or by running the oc get csv command.

The PHASE and conditions values are approximations that are based on available information.

Prerequisites
  • Access to the cluster as a user with the cluster-admin role.

  • Install the OpenShift CLI (oc).

Procedure
  1. Run the following command:

    $ oc get csv
  2. Review the output, checking the PHASE field. For example:

    VERSION    REPLACES                                         PHASE
    4.6.0      performance-addon-operator.v4.5.0                Installing
    4.5.0                                                       Replacing
  3. Run get csv again to verify the output:

    # oc get csv
    Example output
    NAME                                DISPLAY                      VERSION   REPLACES                            PHASE
    performance-addon-operator.v4.5.0   Performance Addon Operator   4.6.0     performance-addon-operator.v4.5.0   Succeeded

Provisioning real-time and low latency workloads

Many industries and organizations need extremely high performance computing and might require low and predictable latency, especially in the financial and telecommunications industries. For these industries, with their unique requirements, OpenShift Container Platform provides a Performance Addon Operator to implement automatic tuning to achieve low latency performance and consistent response time for OpenShift Container Platform applications.

The cluster administrator uses this performance profile configuration that makes it easier to make these changes in a more reliable way. The administrator can specify whether to update the kernel to kernel-rt (real-time), the CPUs that will be reserved for housekeeping, and the CPUs that are used for running the workloads.

The usage of execution probes in conjunction with applications that require guaranteed CPUs can cause latency spikes. It is recommended to use other probes, such as a properly configured set of network probes, as an alternative.

Known limitations for real-time

The RT kernel is only supported on worker nodes.

To fully utilize the real-time mode, the containers must run with elevated privileges. See Set capabilities for a Container for information on granting privileges.

OpenShift Container Platform restricts the allowed capabilities, so you might need to create a SecurityContext as well.

This procedure is fully supported with bare metal installations using Red Hat Enterprise Linux CoreOS (RHCOS) systems.

Establishing the right performance expectations refers to the fact that the real-time kernel is not a panacea. Its objective is consistent, low-latency determinism offering predictable response times. There is some additional kernel overhead associated with the real-time kernel. This is due primarily to handling hardware interruptions in separately scheduled threads. The increased overhead in some workloads results in some degradation in overall throughput. The exact amount of degradation is very workload dependent, ranging from 0% to 30%. However, it is the cost of determinism.

Provisioning a worker with real-time capabilities

  1. Install Performance Addon Operator to the cluster.

  2. Optional: Add a node to the OpenShift Container Platform cluster. See Setting BIOS parameters.

  3. Add the label worker-rt to the worker nodes that require the real-time capability by using the oc command.

  4. Create a new machine config pool for real-time nodes:

    apiVersion: machineconfiguration.openshift.io/v1
    kind: MachineConfigPool
    metadata:
      name: worker-rt
      labels:
        machineconfiguration.openshift.io/role: worker-rt
    spec:
      machineConfigSelector:
        matchExpressions:
          - {
               key: machineconfiguration.openshift.io/role,
               operator: In,
               values: [worker, worker-rt],
            }
      paused: false
      nodeSelector:
        matchLabels:
          node-role.kubernetes.io/worker-rt: ""

    Note that a machine config pool worker-rt is created for group of nodes that have the label worker-rt.

  5. Add the node to the proper machine config pool by using node role labels.

    You must decide which nodes are configured with real-time workloads. You could configure all of the nodes in the cluster, or a subset of the nodes. The Performance Addon Operator that expects all of the nodes are part of a dedicated machine config pool. If you use all of the nodes, you must point the Performance Addon Operator to the worker node role label. If you use a subset, you must group the nodes into a new machine config pool.

  6. Create the PerformanceProfile with the proper set of housekeeping cores and realTimeKernel: enabled: true.

  7. You must set machineConfigPoolSelector in PerformanceProfile:

      apiVersion: performance.openshift.io/v2
      kind: PerformanceProfile
      metadata:
       name: example-performanceprofile
      spec:
      ...
        realTimeKernel:
          enabled: true
        nodeSelector:
           node-role.kubernetes.io/worker-rt: ""
        machineConfigPoolSelector:
           machineconfiguration.openshift.io/role: worker-rt
  8. Verify that a matching machine config pool exists with a label:

    $ oc describe mcp/worker-rt
    Example output
    Name:         worker-rt
    Namespace:
    Labels:       machineconfiguration.openshift.io/role=worker-rt
  9. OpenShift Container Platform will start configuring the nodes, which might involve multiple reboots. Wait for the nodes to settle. This can take a long time depending on the specific hardware you use, but 20 minutes per node is expected.

  10. Verify everything is working as expected.

Verifying the real-time kernel installation

Use this command to verify that the real-time kernel is installed:

$ oc get node -o wide

Note the worker with the role worker-rt that contains the string 4.18.0-211.rt5.23.el8.x86_64:

NAME                               	STATUS   ROLES           	AGE 	VERSION                  	INTERNAL-IP
EXTERNAL-IP   OS-IMAGE                                       	KERNEL-VERSION
CONTAINER-RUNTIME
rt-worker-0.example.com	            Ready	 worker,worker-rt   5d17h   v1.22.1
128.66.135.107   <none>    	        Red Hat Enterprise Linux CoreOS 46.82.202008252340-0 (Ootpa)
4.18.0-211.rt5.23.el8.x86_64   cri-o://1.19.0-90.rhaos4.6.git4a0ac05.el8-rc.1
[...]

Creating a workload that works in real-time

Use the following procedures for preparing a workload that will use real-time capabilities.

Procedure
  1. Create a pod with a QoS class of Guaranteed.

  2. Optional: Disable CPU load balancing for DPDK.

  3. Assign a proper node selector.

When writing your applications, follow the general recommendations described in Application tuning and deployment.

Creating a pod with a QoS class of Guaranteed

Keep the following in mind when you create a pod that is given a QoS class of Guaranteed:

  • Every container in the pod must have a memory limit and a memory request, and they must be the same.

  • Every container in the pod must have a CPU limit and a CPU request, and they must be the same.

The following example shows the configuration file for a pod that has one container. The container has a memory limit and a memory request, both equal to 200 MiB. The container has a CPU limit and a CPU request, both equal to 1 CPU.

apiVersion: v1
kind: Pod
metadata:
  name: qos-demo
  namespace: qos-example
spec:
  containers:
  - name: qos-demo-ctr
    image: <image-pull-spec>
    resources:
      limits:
        memory: "200Mi"
        cpu: "1"
      requests:
        memory: "200Mi"
        cpu: "1"
  1. Create the pod:

    $ oc  apply -f qos-pod.yaml --namespace=qos-example
  2. View detailed information about the pod:

    $ oc get pod qos-demo --namespace=qos-example --output=yaml
    Example output
    spec:
      containers:
        ...
    status:
      qosClass: Guaranteed

    If a container specifies its own memory limit, but does not specify a memory request, OpenShift Container Platform automatically assigns a memory request that matches the limit. Similarly, if a container specifies its own CPU limit, but does not specify a CPU request, OpenShift Container Platform automatically assigns a CPU request that matches the limit.

Optional: Disabling CPU load balancing for DPDK

Functionality to disable or enable CPU load balancing is implemented on the CRI-O level. The code under the CRI-O disables or enables CPU load balancing only when the following requirements are met.

  • The pod must use the performance-<profile-name> runtime class. You can get the proper name by looking at the status of the performance profile, as shown here:

    apiVersion: performance.openshift.io/v1
    kind: PerformanceProfile
    ...
    status:
      ...
      runtimeClass: performance-manual
  • The pod must have the cpu-load-balancing.crio.io: true annotation.

The Performance Addon Operator is responsible for the creation of the high-performance runtime handler config snippet under relevant nodes and for creation of the high-performance runtime class under the cluster. It will have the same content as default runtime handler except it enables the CPU load balancing configuration functionality.

To disable the CPU load balancing for the pod, the Pod specification must include the following fields:

apiVersion: v1
kind: Pod
metadata:
  ...
  annotations:
    ...
    cpu-load-balancing.crio.io: "true"
    ...
  ...
spec:
  ...
  runtimeClassName: performance-<profile_name>
  ...

Only disable CPU load balancing when the CPU manager static policy is enabled and for pods with guaranteed QoS that use whole CPUs. Otherwise, disabling CPU load balancing can affect the performance of other containers in the cluster.

Assigning a proper node selector

The preferred way to assign a pod to nodes is to use the same node selector the performance profile used, as shown here:

apiVersion: v1
kind: Pod
metadata:
  name: example
spec:
  # ...
  nodeSelector:
    node-role.kubernetes.io/worker-rt: ""

Scheduling a workload onto a worker with real-time capabilities

Use label selectors that match the nodes attached to the machine config pool that was configured for low latency by the Performance Addon Operator. For more information, see Assigning pods to nodes.

Configuring huge pages

Nodes must pre-allocate huge pages used in an OpenShift Container Platform cluster. Use the Performance Addon Operator to allocate huge pages on a specific node.

OpenShift Container Platform provides a method for creating and allocating huge pages. Performance Addon Operator provides an easier method for doing this using the performance profile.

For example, in the hugepages pages section of the performance profile, you can specify multiple blocks of size, count, and, optionally, node:

hugepages:
   defaultHugepagesSize: "1G"
   pages:
   - size:  "1G"
     count:  4
     node:  0 (1)
1 node is the NUMA node in which the huge pages are allocated. If you omit node, the pages are evenly spread across all NUMA nodes.

Wait for the relevant machine config pool status that indicates the update is finished.

These are the only configuration steps you need to do to allocate huge pages.

Verification
  • To verify the configuration, see the /proc/meminfo file on the node:

    $ oc debug node/ip-10-0-141-105.ec2.internal
    # grep -i huge /proc/meminfo
    Example output
    AnonHugePages:    ###### ##
    ShmemHugePages:        0 kB
    HugePages_Total:       2
    HugePages_Free:        2
    HugePages_Rsvd:        0
    HugePages_Surp:        0
    Hugepagesize:       #### ##
    Hugetlb:            #### ##
  • Use oc describe to report the new size:

    $ oc describe node worker-0.ocp4poc.example.com | grep -i huge
    Example output
                                       hugepages-1g=true
     hugepages-###:  ###
     hugepages-###:  ###

Allocating multiple huge page sizes

You can request huge pages with different sizes under the same container. This allows you to define more complicated pods consisting of containers with different huge page size needs.

For example, you can define sizes 1G and 2M and the Performance Addon Operator will configure both sizes on the node, as shown here:

spec:
  hugepages:
    defaultHugepagesSize: 1G
    pages:
    - count: 1024
      node: 0
      size: 2M
    - count: 4
      node: 1
      size: 1G

Restricting CPUs for infra and application containers

Generic housekeeping and workload tasks use CPUs in a way that may impact latency-sensitive processes. By default, the container runtime uses all online CPUs to run all containers together, which can result in context switches and spikes in latency. Partitioning the CPUs prevents noisy processes from interfering with latency-sensitive processes by separating them from each other. The following table describes how processes run on a CPU after you have tuned the node using the Performance Addon Operator:

Table 1. Process' CPU assignments
Process type Details

Burstable and BestEffort pods

Runs on any CPU except where low latency workload is running

Infrastructure pods

Runs on any CPU except where low latency workload is running

Interrupts

Redirects to reserved CPUs (optional in OpenShift Container Platform 4.6 and later)

Kernel processes

Pins to reserved CPUs

Latency-sensitive workload pods

Pins to a specific set of exclusive CPUs from the isolated pool

OS processes/systemd services

Pins to reserved CPUs

The allocatable capacity of cores on a node for pods of all QoS process types, Burstable, BestEffort, or Guaranteed, is equal to the capacity of the isolated pool. The capacity of the reserved pool is removed from the node’s total core capacity for use by the cluster and operating system housekeeping duties.

Example 1

A node features a capacity of 100 cores. Using a performance profile, the cluster administrator allocates 50 cores to the isolated pool and 50 cores to the reserved pool. The cluster administrator assigns 25 cores to QoS Guaranteed pods and 25 cores for BestEffort or Burstable pods. This matches the capacity of the isolated pool.

Example 2

A node features a capacity of 100 cores. Using a performance profile, the cluster administrator allocates 50 cores to the isolated pool and 50 cores to the reserved pool. The cluster administrator assigns 50 cores to QoS Guaranteed pods and one core for BestEffort or Burstable pods. This exceeds the capacity of the isolated pool by one core. Pod scheduling fails because of insufficient CPU capacity.

The exact partitioning pattern to use depends on many factors like hardware, workload characteristics and the expected system load. Some sample use cases are as follows:

  • If the latency-sensitive workload uses specific hardware, such as a network interface card (NIC), ensure that the CPUs in the isolated pool are as close as possible to this hardware. At a minimum, you should place the workload in the same Non-Uniform Memory Access (NUMA) node.

  • The reserved pool is used for handling all interrupts. When depending on system networking, allocate a sufficiently-sized reserve pool to handle all the incoming packet interrupts. In 4.6 and later versions, workloads can optionally be labeled as sensitive. The decision regarding which specific CPUs should be used for reserved and isolated partitions requires detailed analysis and measurements. Factors like NUMA affinity of devices and memory play a role. The selection also depends on the workload architecture and the specific use case.

The reserved and isolated CPU pools must not overlap and together must span all available cores in the worker node.

To ensure that housekeeping tasks and workloads do not interfere with each other, specify two groups of CPUs in the spec section of the performance profile.

  • isolated - Specifies the CPUs for the application container workloads. Workloads running on these CPUs experience the lowest latency as well as zero interruptions and can, for example, reach high zero packet loss bandwidth.

  • reserved - Specifies the CPUs for the cluster and operating system housekeeping duties. Threads in the reserved group are often busy. Do not run latency-sensitive applications in the reserved group. Latency-sensitive applications run in the isolated group. .Procedure

    1. Create a performance profile appropriate for the environment’s hardware and topology.

    2. Add the reserved and isolated parameters with the CPUs you want reserved and isolated for the infra and application containers:

      apiVersion: performance.openshift.io/v2
      kind: PerformanceProfile
      metadata:
        name: infra-cpus
      spec:
        cpu:
          reserved: "0-4,9" (1)
          isolated: "5-8" (2)
        nodeSelector: (3)
          node-role.kubernetes.io/worker: ""
      1 Specify which CPUs are for infra containers to perform cluster and operating system housekeeping duties.
      2 Specify which CPUs are for application containers to run workloads.
      3 Specify a node selector to apply the performance profile to specific nodes.

Tuning nodes for low latency with the performance profile

The performance profile lets you control latency tuning aspects of nodes that belong to a certain machine config pool. After you specify your settings, the PerformanceProfile object is compiled into multiple objects that perform the actual node level tuning:

  • A MachineConfig file that manipulates the nodes.

  • A KubeletConfig file that configures the Topology Manager, the CPU Manager, and the OpenShift Container Platform nodes.

  • The Tuned profile that configures the Node Tuning Operator.

Procedure
  1. Prepare a cluster.

  2. Create a machine config pool.

  3. Install the Performance Addon Operator.

  4. Create a performance profile that is appropriate for your hardware and topology. In the performance profile, you can specify whether to update the kernel to kernel-rt, allocation of huge pages, the CPUs that will be reserved for operating system housekeeping processes and CPUs that will be used for running the workloads.

    This is a typical performance profile:

    apiVersion: performance.openshift.io/v1
    kind: PerformanceProfile
    metadata:
     name: performance
    spec:
     cpu:
      isolated: "5-15"
      reserved: "0-4"
     hugepages:
      defaultHugepagesSize: "1G"
      pages:
      -size: "1G"
       count: 16
       node: 0
     realTimeKernel:
      enabled: true  (1)
     numa:  (2)
      topologyPolicy: "best-effort"
     nodeSelector:
      node-role.kubernetes.io/worker-cnf: ""
1 Valid values are true or false. Setting the true value installs the real-time kernel on the node.
2 Use this field to configure the topology manager policy. Valid values are none (default), best-effort, restricted, and single-numa-node. For more information, see Topology Manager Policies.

Performing end-to-end tests for platform verification

The Cloud-native Network Functions (CNF) tests image is a containerized test suite that validates features required to run CNF payloads. You can use this image to validate a CNF-enabled OpenShift cluster where all the components required for running CNF workloads are installed.

The tests run by the image are split into three different phases:

  • Simple cluster validation

  • Setup

  • End to end tests

The validation phase checks that all the features required to be tested are deployed correctly on the cluster.

Validations include:

  • Targeting a machine config pool that belong to the machines to be tested

  • Enabling SCTP on the nodes

  • Having the Performance Addon Operator installed

  • Having the SR-IOV Operator installed

  • Having the PTP Operator installed

  • Using OVN kubernetes as the SDN

The tests need to perform an environment configuration every time they are executed. This involves items such as creating SR-IOV node policies, performance profiles, or PTP profiles. Allowing the tests to configure an already configured cluster might affect the functionality of the cluster. Also, changes to configuration items such as SR-IOV node policy might result in the environment being temporarily unavailable until the configuration change is processed.

Prerequisites

  • The test entrypoint is /usr/bin/test-run.sh. It runs both a setup test set and the real conformance test suite. The minimum requirement is to provide it with a kubeconfig file and its related $KUBECONFIG environment variable, mounted through a volume.

  • The tests assumes that a given feature is already available on the cluster in the form of an Operator, flags enabled on the cluster, or machine configs.

  • Some tests require a pre-existing machine config pool to append their changes to. This must be created on the cluster before running the tests.

    The default worker pool is worker-cnf and can be created with the following manifest:

    apiVersion: machineconfiguration.openshift.io/v1
    kind: MachineConfigPool
    metadata:
      name: worker-cnf
      labels:
        machineconfiguration.openshift.io/role: worker-cnf
    spec:
      machineConfigSelector:
        matchExpressions:
          - {
              key: machineconfiguration.openshift.io/role,
              operator: In,
              values: [worker-cnf, worker],
            }
      paused: false
      nodeSelector:
        matchLabels:
          node-role.kubernetes.io/worker-cnf: ""

    You can use the ROLE_WORKER_CNF variable to override the worker pool name:

    $ docker run -v $(pwd)/:/kubeconfig -e KUBECONFIG=/kubeconfig/kubeconfig -e
    ROLE_WORKER_CNF=custom-worker-pool registry.redhat.io/openshift4/cnf-tests-rhel8:v4.6 /usr/bin/test-run.sh

    Currently, not all tests run selectively on the nodes belonging to the pool.

Running the tests

Assuming the file is in the current folder, the command for running the test suite is:

$ docker run -v $(pwd)/:/kubeconfig -e KUBECONFIG=/kubeconfig/kubeconfig registry.redhat.io/openshift4/cnf-tests-rhel8:v4.6 /usr/bin/test-run.sh

This allows your kubeconfig file to be consumed from inside the running container.

Image parameters

Depending on the requirements, the tests can use different images. There are two images used by the tests that can be changed using the following environment variables:

  • CNF_TESTS_IMAGE

  • DPDK_TESTS_IMAGE

For example, to change the CNF_TESTS_IMAGE with a custom registry run the following command:

$ docker run -v $(pwd)/:/kubeconfig -e KUBECONFIG=/kubeconfig/kubeconfig -e CNF_TESTS_IMAGE="custom-cnf-tests-image:latests" registry.redhat.io/openshift4/cnf-tests-rhel8:v4.6 /usr/bin/test-run.sh

Ginkgo parameters

The test suite is built upon the ginkgo BDD framework. This means that it accepts parameters for filtering or skipping tests.

You can use the -ginkgo.focus parameter to filter a set of tests:

$ docker run -v $(pwd)/:/kubeconfig -e KUBECONFIG=/kubeconfig/kubeconfig registry.redhat.io/openshift4/cnf-tests-rhel8:v4.6 /usr/bin/test-run.sh -ginkgo.focus="performance|sctp"

There is a particular test that requires both SR-IOV and SCTP. Given the selective nature of the focus parameter, this test is triggered by only placing the sriov matcher. If the tests are executed against a cluster where SR-IOV is installed but SCTP is not, adding the -ginkgo.skip=SCTP parameter causes the tests to skip SCTP testing.

Available features

The set of available features to filter are:

  • performance

  • sriov

  • ptp

  • sctp

  • dpdk

Dry run

Use this command to run in dry-run mode. This is useful for checking what is in the test suite and provides output for all of the tests the image would run.

$ docker run -v $(pwd)/:/kubeconfig -e KUBECONFIG=/kubeconfig/kubeconfig registry.redhat.io/openshift4/cnf-tests-rhel8:v4.6 /usr/bin/test-run.sh -ginkgo.dryRun -ginkgo.v

Disconnected mode

The CNF tests image support running tests in a disconnected cluster, meaning a cluster that is not able to reach outer registries. This is done in two steps:

  1. Performing the mirroring.

  2. Instructing the tests to consume the images from a custom registry.

Mirroring the images to a custom registry accessible from the cluster

A mirror executable is shipped in the image to provide the input required by oc to mirror the images needed to run the tests to a local registry.

Run this command from an intermediate machine that has access both to the cluster and to registry.redhat.io over the Internet:

$ docker run -v $(pwd)/:/kubeconfig -e KUBECONFIG=/kubeconfig/kubeconfig registry.redhat.io/openshift4/cnf-tests-rhel8:v4.6 /usr/bin/mirror -registry my.local.registry:5000/ |  oc image mirror -f -

Then, follow the instructions in the following section about overriding the registry used to fetch the images.

Instruct the tests to consume those images from a custom registry

This is done by setting the IMAGE_REGISTRY environment variable:

$ docker run -v $(pwd)/:/kubeconfig -e KUBECONFIG=/kubeconfig/kubeconfig -e IMAGE_REGISTRY="my.local.registry:5000/" -e CNF_TESTS_IMAGE="custom-cnf-tests-image:latests" registry.redhat.io/openshift4/cnf-tests-rhel8:v4.6 /usr/bin/test-run.sh

Mirroring to the cluster internal registry

OpenShift Container Platform provides a built-in container image registry, which runs as a standard workload on the cluster.

Procedure
  1. Gain external access to the registry by exposing it with a route:

    $ oc patch configs.imageregistry.operator.openshift.io/cluster --patch '{"spec":{"defaultRoute":true}}' --type=merge
  2. Fetch the registry endpoint:

    REGISTRY=$(oc get route default-route -n openshift-image-registry --template='{{ .spec.host }}')
  3. Create a namespace for exposing the images:

    $ oc create ns cnftests
  4. Make that image stream available to all the namespaces used for tests. This is required to allow the tests namespaces to fetch the images from the cnftests image stream.

    $ oc policy add-role-to-user system:image-puller system:serviceaccount:sctptest:default --namespace=cnftests
    $ oc policy add-role-to-user system:image-puller system:serviceaccount:cnf-features-testing:default --namespace=cnftests
    $ oc policy add-role-to-user system:image-puller system:serviceaccount:performance-addon-operators-testing:default --namespace=cnftests
    $ oc policy add-role-to-user system:image-puller system:serviceaccount:dpdk-testing:default --namespace=cnftests
    $ oc policy add-role-to-user system:image-puller system:serviceaccount:sriov-conformance-testing:default --namespace=cnftests
  5. Retrieve the docker secret name and auth token:

    SECRET=$(oc -n cnftests get secret | grep builder-docker | awk {'print $1'}
    TOKEN=$(oc -n cnftests get secret $SECRET -o jsonpath="{.data['\.dockercfg']}" | base64 --decode | jq '.["image-registry.openshift-image-registry.svc:5000"].auth')
  6. Write a dockerauth.json similar to this:

    echo "{\"auths\": { \"$REGISTRY\": { \"auth\": $TOKEN } }}" > dockerauth.json
  7. Do the mirroring:

    $ docker run -v $(pwd)/:/kubeconfig -e KUBECONFIG=/kubeconfig/kubeconfig registry.redhat.io/openshift4/cnf-tests-rhel8:v4.6 /usr/bin/mirror -registry $REGISTRY/cnftests |  oc image mirror --insecure=true -a=$(pwd)/dockerauth.json -f -
  8. Run the tests:

    $ docker run -v $(pwd)/:/kubeconfig -e KUBECONFIG=/kubeconfig/kubeconfig -e IMAGE_REGISTRY=image-registry.openshift-image-registry.svc:5000/cnftests cnf-tests-local:latest /usr/bin/test-run.sh

Mirroring a different set of images

Procedure
  1. The mirror command tries to mirror the u/s images by default. This can be overridden by passing a file with the following format to the image:

    [
        {
            "registry": "public.registry.io:5000",
            "image": "imageforcnftests:4.6"
        },
        {
            "registry": "public.registry.io:5000",
            "image": "imagefordpdk:4.6"
        }
    ]
  2. Pass it to the mirror command, for example saving it locally as images.json. With the following command, the local path is mounted in /kubeconfig inside the container and that can be passed to the mirror command.

    $ docker run -v $(pwd)/:/kubeconfig -e KUBECONFIG=/kubeconfig/kubeconfig registry.redhat.io/openshift4/cnf-tests-rhel8:v4.6 /usr/bin/mirror --registry "my.local.registry:5000/" --images "/kubeconfig/images.json" |  oc image mirror -f -

Discovery mode

Discovery mode allows you to validate the functionality of a cluster without altering its configuration. Existing environment configurations are used for the tests. The tests attempt to find the configuration items needed and use those items to execute the tests. If resources needed to run a specific test are not found, the test is skipped, providing an appropriate message to the user. After the tests are finished, no cleanup of the pre-configured configuration items is done, and the test environment can be immediately used for another test run.

Some configuration items are still created by the tests. These are specific items needed for a test to run; for example, a SR-IOV Network. These configuration items are created in custom namespaces and are cleaned up after the tests are executed.

An additional bonus is a reduction in test run times. As the configuration items are already there, no time is needed for environment configuration and stabilization.

To enable discovery mode, the tests must be instructed by setting the DISCOVERY_MODE environment variable as follows:

$ docker run -v $(pwd)/:/kubeconfig:Z -e KUBECONFIG=/kubeconfig/kubeconfig -e
DISCOVERY_MODE=true registry.redhat.io/openshift-kni/cnf-tests /usr/bin/test-run.sh

Required environment configuration prerequisites

SR-IOV tests

Most SR-IOV tests require the following resources:

  • SriovNetworkNodePolicy.

  • At least one with the resource specified by SriovNetworkNodePolicy being allocatable; a resource count of at least 5 is considered sufficient.

Some tests have additional requirements:

  • An unused device on the node with available policy resource, with link state DOWN and not a bridge slave.

  • A SriovNetworkNodePolicy with a MTU value of 9000.

DPDK tests

The DPDK related tests require:

  • A performance profile.

  • A SR-IOV policy.

  • A node with resources available for the SR-IOV policy and available with the PerformanceProfile node selector.

PTP tests
  • A slave PtpConfig (ptp4lOpts="-s" ,phc2sysOpts="-a -r").

  • A node with a label matching the slave PtpConfig.

SCTP tests
  • SriovNetworkNodePolicy.

  • A node matching both the SriovNetworkNodePolicy and a MachineConfig that enables SCTP.

Performance Operator tests

Various tests have different requirements. Some of them are:

  • A performance profile.

  • A performance profile having profile.Spec.CPU.Isolated = 1.

  • A performance profile having profile.Spec.RealTimeKernel.Enabled == true.

  • A node with no huge pages usage.

Limiting the nodes used during tests

The nodes on which the tests are executed can be limited by specifying a NODES_SELECTOR environment variable. Any resources created by the test are then limited to the specified nodes.

$ docker run -v $(pwd)/:/kubeconfig:Z -e KUBECONFIG=/kubeconfig/kubeconfig -e
NODES_SELECTOR=node-role.kubernetes.io/worker-cnf registry.redhat.io/openshift-kni/cnf-tests /usr/bin/test-run.sh

Using a single performance profile

The resources needed by the DPDK tests are higher than those required by the performance test suite. To make the execution faster, the performance profile used by tests can be overridden using one that also serves the DPDK test suite.

To do this, a profile like the following one can be mounted inside the container, and the performance tests can be instructed to deploy it.

apiVersion: performance.openshift.io/v1
kind: PerformanceProfile
metadata:
  name: performance
spec:
  cpu:
    isolated: "4-15"
    reserved: "0-3"
  hugepages:
    defaultHugepagesSize: "1G"
    pages:
    - size: "1G"
      count: 16
      node: 0
  realTimeKernel:
    enabled: true
  nodeSelector:
    node-role.kubernetes.io/worker-cnf: ""

To override the performance profile used, the manifest must be mounted inside the container and the tests must be instructed by setting the PERFORMANCE_PROFILE_MANIFEST_OVERRIDE parameter as follows:

$ docker run -v $(pwd)/:/kubeconfig:Z -e KUBECONFIG=/kubeconfig/kubeconfig -e
PERFORMANCE_PROFILE_MANIFEST_OVERRIDE=/kubeconfig/manifest.yaml registry.redhat.io/openshift-kni/cnf-tests /usr/bin/test-run.sh

Disabling the performance profile cleanup

When not running in discovery mode, the suite cleans up all the created artifacts and configurations. This includes the performance profile.

When deleting the performance profile, the machine config pool is modified and nodes are rebooted. After a new iteration, a new profile is created. This causes long test cycles between runs.

To speed up this process, set CLEAN_PERFORMANCE_PROFILE="false" to instruct the tests not to clean the performance profile. In this way, the next iteration will not need to create it and wait for it to be applied.

$ docker run -v $(pwd)/:/kubeconfig:Z -e KUBECONFIG=/kubeconfig/kubeconfig -e
CLEAN_PERFORMANCE_PROFILE="false" registry.redhat.io/openshift-kni/cnf-tests /usr/bin/test-run.sh

Troubleshooting

The cluster must be reached from within the container. You can verify this by running:

$ docker run -v $(pwd)/:/kubeconfig -e KUBECONFIG=/kubeconfig/kubeconfig
registry.redhat.io/openshift-kni/cnf-tests oc get nodes

If this does not work, it could be caused by spanning across DNS, MTU size, or firewall issues.

Test reports

CNF end-to-end tests produce two outputs: a JUnit test output and a test failure report.

JUnit test output

A JUnit-compliant XML is produced by passing the --junit parameter together with the path where the report is dumped:

$ docker run -v $(pwd)/:/kubeconfig -v $(pwd)/junitdest:/path/to/junit -e KUBECONFIG=/kubeconfig/kubeconfig registry.redhat.io/openshift4/cnf-tests-rhel8:v4.6 /usr/bin/test-run.sh --junit /path/to/junit

Test failure report

A report with information about the cluster state and resources for troubleshooting can be produced by passing the --report parameter with the path where the report is dumped:

$ docker run -v $(pwd)/:/kubeconfig -v $(pwd)/reportdest:/path/to/report -e KUBECONFIG=/kubeconfig/kubeconfig registry.redhat.io/openshift4/cnf-tests-rhel8:v4.6 /usr/bin/test-run.sh --report /path/to/report

A note on podman

When executing podman as non root and non privileged, mounting paths can fail with "permission denied" errors. To make it work, append :Z to the volumes creation; for example, -v $(pwd)/:/kubeconfig:Z to allow podman to do the proper SELinux relabeling.

Running on OpenShift Container Platform 4.4

With the exception of the following, the CNF end-to-end tests are compatible with OpenShift Container Platform 4.4:

[test_id:28466][crit:high][vendor:cnf-qe@redhat.com][level:acceptance] Should contain configuration injected through openshift-node-performance profile
[test_id:28467][crit:high][vendor:cnf-qe@redhat.com][level:acceptance] Should contain configuration injected through the openshift-node-performance profile

You can skip these tests by adding the -ginkgo.skip “28466|28467" parameter.

Using a single performance profile

The DPDK tests require more resources than what is required by the performance test suite. To make the execution faster, you can override the performance profile used by the tests using a profile that also serves the DPDK test suite.

To do this, use a profile like the following one that can be mounted inside the container, and the performance tests can be instructed to deploy it.

apiVersion: performance.openshift.io/v1
kind: PerformanceProfile
metadata:
 name: performance
spec:
 cpu:
  isolated: "5-15"
  reserved: "0-4"
 hugepages:
  defaultHugepagesSize: "1G"
  pages:
  -size: "1G"
   count: 16
   node: 0
 realTimeKernel:
  enabled: true
 numa:
  topologyPolicy: "best-effort"
 nodeSelector:
  node-role.kubernetes.io/worker-cnf: ""

To override the performance profile, the manifest must be mounted inside the container and the tests must be instructed by setting the PERFORMANCE_PROFILE_MANIFEST_OVERRIDE:

$ docker run -v $(pwd)/:/kubeconfig:Z -e KUBECONFIG=/kubeconfig/kubeconfig -e PERFORMANCE_PROFILE_MANIFEST_OVERRIDE=/kubeconfig/manifest.yaml registry.redhat.io/openshift4/cnf-tests-rhel8:v4.6 /usr/bin/test-run.sh

Impacts on the cluster

Depending on the feature, running the test suite could cause different impacts on the cluster. In general, only the SCTP tests do not change the cluster configuration. All of the other features have various impacts on the configuration.

SCTP

SCTP tests just run different pods on different nodes to check connectivity. The impacts on the cluster are related to running simple pods on two nodes.

SR-IOV

SR-IOV tests require changes in the SR-IOV network configuration, where the tests create and destroy different types of configuration.

This might have an impact if existing SR-IOV network configurations are already installed on the cluster, because there may be conflicts depending on the priority of such configurations.

At the same time, the result of the tests might be affected by existing configurations.

PTP

PTP tests apply a PTP configuration to a set of nodes of the cluster. As with SR-IOV, this might conflict with any existing PTP configuration already in place, with unpredictable results.

Performance

Performance tests apply a performance profile to the cluster. The effect of this is changes in the node configuration, reserving CPUs, allocating memory huge pages, and setting the kernel packages to be realtime. If an existing profile named performance is already available on the cluster, the tests do not deploy it.

DPDK

DPDK relies on both performance and SR-IOV features, so the test suite configures both a performance profile and SR-IOV networks, so the impacts are the same as those described in SR-IOV testing and performance testing.

Cleaning up

After running the test suite, all the dangling resources are cleaned up.

Debugging low latency CNF tuning status

The PerformanceProfile custom resource (CR) contains status fields for reporting tuning status and debugging latency degradation issues. These fields report on conditions that describe the state of the operator’s reconciliation functionality.

A typical issue can arise when the status of machine config pools that are attached to the performance profile are in a degraded state, causing the PerformanceProfile status to degrade. In this case, the machine config pool issues a failure message.

The Performance Addon Operator contains the performanceProfile.spec.status.Conditions status field:

Status:
  Conditions:
    Last Heartbeat Time:   2020-06-02T10:01:24Z
    Last Transition Time:  2020-06-02T10:01:24Z
    Status:                True
    Type:                  Available
    Last Heartbeat Time:   2020-06-02T10:01:24Z
    Last Transition Time:  2020-06-02T10:01:24Z
    Status:                True
    Type:                  Upgradeable
    Last Heartbeat Time:   2020-06-02T10:01:24Z
    Last Transition Time:  2020-06-02T10:01:24Z
    Status:                False
    Type:                  Progressing
    Last Heartbeat Time:   2020-06-02T10:01:24Z
    Last Transition Time:  2020-06-02T10:01:24Z
    Status:                False
    Type:                  Degraded

The Status field contains Conditions that specify Type values that indicate the status of the performance profile:

Available

All machine configs and Tuned profiles have been created successfully and are available for cluster components are responsible to process them (NTO, MCO, Kubelet).

Upgradeable

Indicates whether the resources maintained by the Operator are in a state that is safe to upgrade.

Progressing

Indicates that the deployment process from the performance profile has started.

Degraded

Indicates an error if:

  • Validation of the performance profile has failed.

  • Creation of all relevant components did not complete successfully.

Each of these types contain the following fields:

Status

The state for the specific type (true or false).

Timestamp

The transaction timestamp.

Reason string

The machine readable reason.

Message string

The human readable reason describing the state and error details, if any.

Machine config pools

A performance profile and its created products are applied to a node according to an associated machine config pool (MCP). The MCP holds valuable information about the progress of applying the machine configurations created by performance addons that encompass kernel args, kube config, huge pages allocation, and deployment of rt-kernel. The performance addons controller monitors changes in the MCP and updates the performance profile status accordingly.

The only conditions returned by the MCP to the performance profile status is when the MCP is Degraded, which leads to performaceProfile.status.condition.Degraded = true.

Example

The following example is for a performance profile with an associated machine config pool (worker-cnf) that was created for it:

  1. The associated machine config pool is in a degraded state:

    # oc get mcp
    Example output
    NAME         CONFIG                                                 UPDATED   UPDATING   DEGRADED   MACHINECOUNT   READYMACHINECOUNT   UPDATEDMACHINECOUNT   DEGRADEDMACHINECOUNT   AGE
    master       rendered-master-2ee57a93fa6c9181b546ca46e1571d2d       True      False      False      3              3                   3                     0                      2d21h
    worker       rendered-worker-d6b2bdc07d9f5a59a6b68950acf25e5f       True      False      False      2              2                   2                     0                      2d21h
    worker-cnf   rendered-worker-cnf-6c838641b8a08fff08dbd8b02fb63f7c   False     True       True       2              1                   1                     1                      2d20h
  2. The describe section of the MCP shows the reason:

    # oc describe mcp worker-cnf
    Example output
      Message:               Node node-worker-cnf is reporting: "prepping update:
      machineconfig.machineconfiguration.openshift.io \"rendered-worker-cnf-40b9996919c08e335f3ff230ce1d170\" not
      found"
        Reason:                1 nodes are reporting degraded status on sync
  3. The degraded state should also appear under the performance profile status field marked as degraded = true:

    # oc describe performanceprofiles performance
    Example output
    Message: Machine config pool worker-cnf Degraded Reason: 1 nodes are reporting degraded status on sync.
    Machine config pool worker-cnf Degraded Message: Node yquinn-q8s5v-w-b-z5lqn.c.openshift-gce-devel.internal is
    reporting: "prepping update: machineconfig.machineconfiguration.openshift.io
    \"rendered-worker-cnf-40b9996919c08e335f3ff230ce1d170\" not found".    Reason:  MCPDegraded
       Status:  True
       Type:    Degraded

Collecting low latency tuning debugging data for Red Hat Support

When opening a support case, it is helpful to provide debugging information about your cluster to Red Hat Support.

The must-gather tool enables you to collect diagnostic information about your OpenShift Container Platform cluster, including node tuning, NUMA topology, and other information needed to debug issues with low latency setup.

For prompt support, supply diagnostic information for both OpenShift Container Platform and low latency tuning.

About the must-gather tool

The oc adm must-gather CLI command collects the information from your cluster that is most likely needed for debugging issues, such as:

  • Resource definitions

  • Audit logs

  • Service logs

You can specify one or more images when you run the command by including the --image argument. When you specify an image, the tool collects data related to that feature or product. When you run oc adm must-gather, a new pod is created on the cluster. The data is collected on that pod and saved in a new directory that starts with must-gather.local. This directory is created in your current working directory.

About collecting low latency tuning data

Use the oc adm must-gather CLI command to collect information about your cluster, including features and objects associated with low latency tuning, including:

  • The Performance Addon Operator namespaces and child objects.

  • MachineConfigPool and associated MachineConfig objects.

  • The Node Tuning Operator and associated Tuned objects.

  • Linux Kernel command line options.

  • CPU and NUMA topology

  • Basic PCI device information and NUMA locality.

To collect Performance Addon Operator debugging information with must-gather, you must specify the Performance Addon Operator must-gather image:

--image=registry.redhat.io/openshift4/performance-addon-operator-must-gather-rhel8:v4.6.

Gathering data about specific features

You can gather debugging information about specific features by using the oc adm must-gather CLI command with the --image or --image-stream argument. The must-gather tool supports multiple images, so you can gather data about more than one feature by running a single command.

To collect the default must-gather data in addition to specific feature data, add the --image-stream=openshift/must-gather argument.

Prerequisites
  • Access to the cluster as a user with the cluster-admin role.

  • The OpenShift Container Platform CLI (oc) installed.

Procedure
  1. Navigate to the directory where you want to store the must-gather data.

  2. Run the oc adm must-gather command with one or more --image or --image-stream arguments. For example, the following command gathers both the default cluster data and information specific to the Performance Addon Operator:

    $ oc adm must-gather \
     --image-stream=openshift/must-gather \ (1)
    
     --image=registry.redhat.io/openshift4/performance-addon-operator-must-gather-rhel8:v4.6 (2)
    
    1 The default OpenShift Container Platform must-gather image.
    2 The must-gather image for low latency tuning diagnostics.
  3. Create a compressed file from the must-gather directory that was created in your working directory. For example, on a computer that uses a Linux operating system, run the following command:

     $ tar cvaf must-gather.tar.gz must-gather.local.5421342344627712289/ (1)
    1 Replace must-gather-local.5421342344627712289/ with the actual directory name.
  4. Attach the compressed file to your support case on the Red Hat Customer Portal.

Additional resources