Pod Priority and Preemption
FEATURE STATE: Kubernetes v1.8 alpha
This feature is currently in a alpha state, meaning:
- The version names contain alpha (e.g. v1alpha1).
- Might be buggy. Enabling the feature may expose bugs. Disabled by default.
- Support for feature may be dropped at any time without notice.
- The API may change in incompatible ways in a later software release without notice.
- Recommended for use only in short-lived testing clusters, due to increased risk of bugs and lack of long-term support.
Pods in Kubernetes 1.8 and later can have priority. Priority indicates the importance of a Pod relative to other Pods. When a Pod cannot be scheduled, the scheduler tries to preempt (evict) lower priority Pods to make scheduling of the pending Pod possible. In a future Kubernetes release, priority will also affect out-of-resource eviction ordering on the Node.
Note: Preemption does not respect PodDisruptionBudget; see the limitations section for more details.
- How to use priority and preemption
- Enabling priority and preemption
- PriorityClass
- Pod priority
- Preemption
How to use priority and preemption
To use priority and preemption in Kubernetes 1.8, follow these steps:
-
Enable the feature.
-
Add one or more PriorityClasses.
-
Create Pods with
PriorityClassNameset to one of the added PriorityClasses. Of course you do not need to create the Pods directly; normally you would addPriorityClassNameto the Pod template of a collection object like a Deployment.
The following sections provide more information about these steps.
Enabling priority and preemption
Pod priority and preemption is disabled by default in Kubernetes 1.8. To enable the feature, set this command-line flag for the API server, scheduler and kubelet:
--feature-gates=PodPriority=true
Also enable scheduling.k8s.io/v1alpha1 API and Priority admission controller in API server:
--runtime-config=scheduling.k8s.io/v1alpha1=true --admission-control=Controller-Foo,Controller-Bar,...,Priority
After the feature is enabled, you can create PriorityClasses
and create Pods with PriorityClassName set.
If you try the feature and then decide to disable it, you must remove the PodPriority command-line flag or set it to false, and then restart the API server and scheduler. After the feature is disabled, the existing Pods keep their priority fields, but preemption is disabled, and priority fields are ignored, and you cannot set PriorityClassName in new Pods.
PriorityClass
A PriorityClass is a non-namespaced object that defines a mapping from a priority
class name to the integer value of the priority. The name is specified in the name
field of the PriorityClass object’s metadata. The value is specified in the required
value field. The higher the value, the higher the priority.
A PriorityClass object can have any 32-bit integer value smaller than or equal to 1 billion. Larger numbers are reserved for critical system Pods that should not normally be preempted or evicted. A cluster admin should create one PriorityClass object for each such mapping that they want.
PriorityClass also has two optional fields: globalDefault and description.
The globalDefault field indicates that the value of this PriorityClass should
be used for Pods without a PriorityClassName. Only one PriorityClass with
globalDefault set to true can exist in the system. If there is no PriorityClass
with globalDefault set, the priority of Pods with no PriorityClassName is zero.
The description field is an arbitrary string. It is meant to tell users of
the cluster when they should use this PriorityClass.
Note 1: If you upgrade your existing cluster and enable this feature, the priority of your existing Pods will be considered to be zero.
Note 2: Addition of a PriorityClass with globalDefault set to true does not
change the priorities of existing Pods. The value of such a PriorityClass is used only
for Pods created after the PriorityClass is added.
Note 3: If you delete a PriorityClass, existing Pods that use the name of the deleted priority class remain unchanged, but you are not able to create more Pods that use the name of the deleted PriorityClass.
Example PriorityClass
apiVersion: scheduling.k8s.io/v1alpha1
kind: PriorityClass
metadata:
name: high-priority
value: 1000000
globalDefault: false
description: "This priority class should be used for XYZ service pods only."
Pod priority
After you have one or more PriorityClasses, you can create Pods that specify one
of those PriorityClass names in their specifications. The priority admission
controller uses the priorityClassName field and populates the integer value
of the priority. If the priority class is not found, the Pod is rejected.
The following YAML is an example of a Pod configuration that uses the PriorityClass created in the preceding example. The priority admission controller checks the specification and resolves the priority of the Pod to 1000000.
apiVersion: v1
kind: Pod
metadata:
name: nginx
labels:
env: test
spec:
containers:
- name: nginx
image: nginx
imagePullPolicy: IfNotPresent
priorityClassName: high-priority
Preemption
When Pods are created, they go to a queue and wait to be scheduled. The scheduler picks a Pod from the queue and tries to schedule it on a Node. If no Node is found that satisfies all the specified requirements of the Pod, preemption logic is triggered for the pending Pod. Let’s call the pending pod P. Preemption logic tries to find a Node where removal of one or more Pods with lower priority than P would enable P to be scheduled on that Node. If such a Node is found, one or more lower priority Pods get deleted from the Node. After the Pods are gone, P can be scheduled on the Node.
Limitations of preemption (alpha version)
Starvation of preempting Pod
When Pods are preempted, the victims get their graceful termination period. They have that much time to finish their work and exit. If they don’t, they are killed. This graceful termination period creates a time gap between the point that the scheduler preempts Pods and the time when the pending Pod (P) can be scheduled on the Node (N). In the meantime, the scheduler keeps scheduling other pending Pods. As victims exit or get terminated, the scheduler tries to schedule Pods in the pending queue, and one or more of them may be considered and scheduled to N before the scheduler considers scheduling P on N. In such a case, it is likely that when all the victims exit, Pod P won’t fit on Node N anymore. So, scheduler will have to preempt other Pods on Node N or another Node so that P can be scheduled. This scenario might be repeated again for the second and subsequent rounds of preemption, and P might not get scheduled for a while. This scenario can cause problems in various clusters, but is particularly problematic in clusters with a high Pod creation rate.
We will address this problem in the beta version of Pod preemption. The solution we plan to implement is provided here.
PodDisruptionBudget is not supported
A Pod Disruption Budget (PDB) allows application owners to limit the number Pods of a replicated application that are down simultaneously from voluntary disruptions. However, the alpha version of preemption does not respect PDB when choosing preemption victims. We plan to add PDB support in beta, but even in beta, respecting PDB will be best effort. The Scheduler will try to find victims whose PDB won’t be violated by preemption, but if no such victims are found, preemption will still happen, and lower priority Pods will be removed despite their PDBs being violated.
Inter-Pod affinity on lower-priority Pods
In version 1.8, a Node is considered for preemption only when the answer to this question is yes: “If all the Pods with lower priority than the pending Pod are removed from the Node, can the pending pod be scheduled on the Node?”
Note: Preemption does not necessarily remove all lower-priority Pods. If the pending pod can be scheduled by removing fewer than all lower-priority Pods, then only a portion of the lower-priority Pods are removed. Even so, the answer to the preceding question must be yes. If the answer is no, the Node is not considered for preemption.
If a pending Pod has inter-pod affinity to one or more of the lower-priority Pods on the Node, the inter-Pod affinity rule cannot be satisfied in the absence of those lower-priority Pods. In this case, the scheduler does not preempt any Pods on the Node. Instead, it looks for another Node. The scheduler might find a suitable Node or it might not. There is no guarantee that the pending Pod can be scheduled.
We might address this issue in future versions, but we don’t have a clear plan yet. We will not consider it a blocker for Beta or GA. Part of the reason is that finding the set of lower-priority Pods that satisfy all inter-Pod affinity rules is computationally expensive, and adds substantial complexity to the preemption logic. Besides, even if preemption keeps the lower-priority Pods to satisfy inter-Pod affinity, the lower priority Pods might be preempted later by other Pods, which removes the benefits of having the complex logic of respecting inter-Pod affinity.
Our recommended solution for this problem is to create inter-Pod affinity only towards equal or higher priority pods.
Cross node preemption
Suppose a Node N is being considered for preemption so that a pending Pod P can be scheduled on N. P might become feasible on N only if a Pod on another Node is preempted. Here’s an example:
- Pod P is being considered for Node N.
- Pod Q is running on another Node in the same zone as Node N.
- Pod P has anit-affinity with Pod Q.
- There are no other cases of anti-affinity between Pod P and other Pods in the zone.
- In order to schedule Pod P on Node N, Pod Q should be preempted, but scheduler does not perform cross-node preemption. So, Pod P will be deemed unschedulable on Node N.
If Pod Q were removed from its Node, the anti-affinity violation would be gone, and Pod P could possibly be scheduled on Node N.
We may consider adding cross Node preemption in future versions if we find an algorithm with reasonable performance. We cannot promise anything at this point, and cross Node preemption will not be considered a blocker for Beta or GA.