Using the Node Tuning Operator

Learn about the Node Tuning Operator and how you can use it to manage node-level tuning by orchestrating the tuned daemon.

Purpose

The Node Tuning Operator helps you manage node-level tuning by orchestrating the TuneD daemon and achieves low latency performance by using the Performance Profile controller. The majority of high-performance applications require some level of kernel tuning. The Node Tuning Operator provides a unified management interface to users of node-level sysctls and more flexibility to add custom tuning specified by user needs.

The Operator manages the containerized TuneD daemon for OKD as a Kubernetes daemon set. It ensures the custom tuning specification is passed to all containerized TuneD daemons running in the cluster in the format that the daemons understand. The daemons run on all nodes in the cluster, one per node.

Node-level settings applied by the containerized TuneD daemon are rolled back on an event that triggers a profile change or when the containerized TuneD daemon is terminated gracefully by receiving and handling a termination signal.

The Node Tuning Operator uses the Performance Profile controller to implement automatic tuning to achieve low latency performance for OKD applications.

The cluster administrator configures a performance profile to define node-level settings such as the following:

  • Updating the kernel to kernel-rt.

  • Choosing CPUs for housekeeping.

  • Choosing CPUs for running workloads.

Currently, disabling CPU load balancing is not supported by cgroup v2. As a result, you might not get the desired behavior from performance profiles if you have cgroup v2 enabled. Enabling cgroup v2 is not recommended if you are using performance profiles.

The Node Tuning Operator is part of a standard OKD installation in version 4.1 and later.

In earlier versions of OKD, the Performance Addon Operator was used to implement automatic tuning to achieve low latency performance for OpenShift applications. In OKD 4.11 and later, this functionality is part of the Node Tuning Operator.

Accessing an example Node Tuning Operator specification

Use this process to access an example Node Tuning Operator specification.

Procedure

  • Run the following command to access an example Node Tuning Operator specification:

    1. oc get tuned.tuned.openshift.io/default -o yaml -n openshift-cluster-node-tuning-operator

The default CR is meant for delivering standard node-level tuning for the OKD platform and it can only be modified to set the Operator Management state. Any other custom changes to the default CR will be overwritten by the Operator. For custom tuning, create your own Tuned CRs. Newly created CRs will be combined with the default CR and custom tuning applied to OKD nodes based on node or pod labels and profile priorities.

While in certain situations the support for pod labels can be a convenient way of automatically delivering required tuning, this practice is discouraged and strongly advised against, especially in large-scale clusters. The default Tuned CR ships without pod label matching. If a custom profile is created with pod label matching, then the functionality will be enabled at that time. The pod label functionality will be deprecated in future versions of the Node Tuning Operator.

Custom tuning specification

The custom resource (CR) for the Operator has two major sections. The first section, profile:, is a list of TuneD profiles and their names. The second, recommend:, defines the profile selection logic.

Multiple custom tuning specifications can co-exist as multiple CRs in the Operator’s namespace. The existence of new CRs or the deletion of old CRs is detected by the Operator. All existing custom tuning specifications are merged and appropriate objects for the containerized TuneD daemons are updated.

Management state

The Operator Management state is set by adjusting the default Tuned CR. By default, the Operator is in the Managed state and the spec.managementState field is not present in the default Tuned CR. Valid values for the Operator Management state are as follows:

  • Managed: the Operator will update its operands as configuration resources are updated

  • Unmanaged: the Operator will ignore changes to the configuration resources

  • Removed: the Operator will remove its operands and resources the Operator provisioned

Profile data

The profile: section lists TuneD profiles and their names.

  1. profile:
  2. - name: tuned_profile_1
  3. data: |
  4. # TuneD profile specification
  5. [main]
  6. summary=Description of tuned_profile_1 profile
  7. [sysctl]
  8. net.ipv4.ip_forward=1
  9. # ... other sysctl's or other TuneD daemon plugins supported by the containerized TuneD
  10. # ...
  11. - name: tuned_profile_n
  12. data: |
  13. # TuneD profile specification
  14. [main]
  15. summary=Description of tuned_profile_n profile
  16. # tuned_profile_n profile settings

Recommended profiles

The profile: selection logic is defined by the recommend: section of the CR. The recommend: section is a list of items to recommend the profiles based on a selection criteria.

  1. recommend:
  2. <recommend-item-1>
  3. # ...
  4. <recommend-item-n>

The individual items of the list:

  1. - machineConfigLabels: (1)
  2. <mcLabels> (2)
  3. match: (3)
  4. <match> (4)
  5. priority: <priority> (5)
  6. profile: <tuned_profile_name> (6)
  7. operand: (7)
  8. debug: <bool> (8)
  9. tunedConfig:
  10. reapply_sysctl: <bool> (9)
1Optional.
2A dictionary of key/value MachineConfig labels. The keys must be unique.
3If omitted, profile match is assumed unless a profile with a higher priority matches first or machineConfigLabels is set.
4An optional list.
5Profile ordering priority. Lower numbers mean higher priority (0 is the highest priority).
6A TuneD profile to apply on a match. For example tuned_profile_1.
7Optional operand configuration.
8Turn debugging on or off for the TuneD daemon. Options are true for on or false for off. The default is false.
9Turn reapply_sysctl functionality on or off for the TuneD daemon. Options are true for on and false for off.

<match> is an optional list recursively defined as follows:

  1. - label: <label_name> (1)
  2. value: <label_value> (2)
  3. type: <label_type> (3)
  4. <match> (4)
1Node or pod label name.
2Optional node or pod label value. If omitted, the presence of <label_name> is enough to match.
3Optional object type (node or pod). If omitted, node is assumed.
4An optional <match> list.

If <match> is not omitted, all nested <match> sections must also evaluate to true. Otherwise, false is assumed and the profile with the respective <match> section will not be applied or recommended. Therefore, the nesting (child <match> sections) works as logical AND operator. Conversely, if any item of the <match> list matches, the entire <match> list evaluates to true. Therefore, the list acts as logical OR operator.

If machineConfigLabels is defined, machine config pool based matching is turned on for the given recommend: list item. <mcLabels> specifies the labels for a machine config. The machine config is created automatically to apply host settings, such as kernel boot parameters, for the profile <tuned_profile_name>. This involves finding all machine config pools with machine config selector matching <mcLabels> and setting the profile <tuned_profile_name> on all nodes that are assigned the found machine config pools. To target nodes that have both master and worker roles, you must use the master role.

The list items match and machineConfigLabels are connected by the logical OR operator. The match item is evaluated first in a short-circuit manner. Therefore, if it evaluates to true, the machineConfigLabels item is not considered.

When using machine config pool based matching, it is advised to group nodes with the same hardware configuration into the same machine config pool. Not following this practice might result in TuneD operands calculating conflicting kernel parameters for two or more nodes sharing the same machine config pool.

Example: node or pod label based matching

  1. - match:
  2. - label: tuned.openshift.io/elasticsearch
  3. match:
  4. - label: node-role.kubernetes.io/master
  5. - label: node-role.kubernetes.io/infra
  6. type: pod
  7. priority: 10
  8. profile: openshift-control-plane-es
  9. - match:
  10. - label: node-role.kubernetes.io/master
  11. - label: node-role.kubernetes.io/infra
  12. priority: 20
  13. profile: openshift-control-plane
  14. - priority: 30
  15. profile: openshift-node

The CR above is translated for the containerized TuneD daemon into its recommend.conf file based on the profile priorities. The profile with the highest priority (10) is openshift-control-plane-es and, therefore, it is considered first. The containerized TuneD daemon running on a given node looks to see if there is a pod running on the same node with the tuned.openshift.io/elasticsearch label set. If not, the entire <match> section evaluates as false. If there is such a pod with the label, in order for the <match> section to evaluate to true, the node label also needs to be node-role.kubernetes.io/master or node-role.kubernetes.io/infra.

If the labels for the profile with priority 10 matched, openshift-control-plane-es profile is applied and no other profile is considered. If the node/pod label combination did not match, the second highest priority profile (openshift-control-plane) is considered. This profile is applied if the containerized TuneD pod runs on a node with labels node-role.kubernetes.io/master or node-role.kubernetes.io/infra.

Finally, the profile openshift-node has the lowest priority of 30. It lacks the <match> section and, therefore, will always match. It acts as a profile catch-all to set openshift-node profile, if no other profile with higher priority matches on a given node.

Decision workflow

Example: machine config pool based matching

  1. apiVersion: tuned.openshift.io/v1
  2. kind: Tuned
  3. metadata:
  4. name: openshift-node-custom
  5. namespace: openshift-cluster-node-tuning-operator
  6. spec:
  7. profile:
  8. - data: |
  9. [main]
  10. summary=Custom OpenShift node profile with an additional kernel parameter
  11. include=openshift-node
  12. [bootloader]
  13. cmdline_openshift_node_custom=+skew_tick=1
  14. name: openshift-node-custom
  15. recommend:
  16. - machineConfigLabels:
  17. machineconfiguration.openshift.io/role: "worker-custom"
  18. priority: 20
  19. profile: openshift-node-custom

To minimize node reboots, label the target nodes with a label the machine config pool’s node selector will match, then create the Tuned CR above and finally create the custom machine config pool itself.

Cloud provider-specific TuneD profiles

With this functionality, all Cloud provider-specific nodes can conveniently be assigned a TuneD profile specifically tailored to a given Cloud provider on a OKD cluster. This can be accomplished without adding additional node labels or grouping nodes into machine config pools.

This functionality takes advantage of spec.providerID node object values in the form of <cloud-provider>://<cloud-provider-specific-id> and writes the file /var/lib/tuned/provider with the value <cloud-provider> in NTO operand containers. The content of this file is then used by TuneD to load provider-<cloud-provider> profile if such profile exists.

The openshift profile that both openshift-control-plane and openshift-node profiles inherit settings from is now updated to use this functionality through the use of conditional profile loading. Neither NTO nor TuneD currently include any Cloud provider-specific profiles. However, it is possible to create a custom profile provider-<cloud-provider> that will be applied to all Cloud provider-specific cluster nodes.

Example GCE Cloud provider profile

  1. apiVersion: tuned.openshift.io/v1
  2. kind: Tuned
  3. metadata:
  4. name: provider-gce
  5. namespace: openshift-cluster-node-tuning-operator
  6. spec:
  7. profile:
  8. - data: |
  9. [main]
  10. summary=GCE Cloud provider-specific profile
  11. # Your tuning for GCE Cloud provider goes here.
  12. name: provider-gce

Due to profile inheritance, any setting specified in the provider-<cloud-provider> profile will be overwritten by the openshift profile and its child profiles.

Default profiles set on a cluster

The following are the default profiles set on a cluster.

  1. apiVersion: tuned.openshift.io/v1
  2. kind: Tuned
  3. metadata:
  4. name: default
  5. namespace: openshift-cluster-node-tuning-operator
  6. spec:
  7. profile:
  8. - data: |
  9. [main]
  10. summary=Optimize systems running OpenShift (provider specific parent profile)
  11. include=-provider-${f:exec:cat:/var/lib/tuned/provider},openshift
  12. name: openshift
  13. recommend:
  14. - profile: openshift-control-plane
  15. priority: 30
  16. match:
  17. - label: node-role.kubernetes.io/master
  18. - label: node-role.kubernetes.io/infra
  19. - profile: openshift-node
  20. priority: 40

Starting with OKD 4.9, all OpenShift TuneD profiles are shipped with the TuneD package. You can use the oc exec command to view the contents of these profiles:

  1. $ oc exec $tuned_pod -n openshift-cluster-node-tuning-operator -- find /usr/lib/tuned/openshift{,-control-plane,-node} -name tuned.conf -exec grep -H ^ {} \;

Supported TuneD daemon plugins

Excluding the [main] section, the following TuneD plugins are supported when using custom profiles defined in the profile: section of the Tuned CR:

  • audio

  • cpu

  • disk

  • eeepc_she

  • modules

  • mounts

  • net

  • scheduler

  • scsi_host

  • selinux

  • sysctl

  • sysfs

  • usb

  • video

  • vm

  • bootloader

There is some dynamic tuning functionality provided by some of these plugins that is not supported. The following TuneD plugins are currently not supported:

  • script

  • systemd

The TuneD bootloader plugin only supports Fedora CoreOS (FCOS) worker nodes.

Additional resources