@@ -5820,6 +5820,189 @@ spec:
58205820 type: object
58215821 type: array
58225822 x-kubernetes-list-type: atomic
5823+ topologySpreadConstraints:
5824+ description: |-
5825+ TopologySpreadConstraints describes how a group of pods ought to spread across topology
5826+ domains. Scheduler will schedule pods in a way which abides by the constraints.
5827+ All topologySpreadConstraints are ANDed.
5828+ items:
5829+ description: TopologySpreadConstraint specifies how to spread
5830+ matching pods among the given topology.
5831+ properties:
5832+ labelSelector:
5833+ description: |-
5834+ LabelSelector is used to find matching pods.
5835+ Pods that match this label selector are counted to determine the number of pods
5836+ in their corresponding topology domain.
5837+ properties:
5838+ matchExpressions:
5839+ description: matchExpressions is a list of label selector
5840+ requirements. The requirements are ANDed.
5841+ items:
5842+ description: |-
5843+ A label selector requirement is a selector that contains values, a key, and an operator that
5844+ relates the key and values.
5845+ properties:
5846+ key:
5847+ description: key is the label key that the selector
5848+ applies to.
5849+ type: string
5850+ operator:
5851+ description: |-
5852+ operator represents a key's relationship to a set of values.
5853+ Valid operators are In, NotIn, Exists and DoesNotExist.
5854+ type: string
5855+ values:
5856+ description: |-
5857+ values is an array of string values. If the operator is In or NotIn,
5858+ the values array must be non-empty. If the operator is Exists or DoesNotExist,
5859+ the values array must be empty. This array is replaced during a strategic
5860+ merge patch.
5861+ items:
5862+ type: string
5863+ type: array
5864+ x-kubernetes-list-type: atomic
5865+ required:
5866+ - key
5867+ - operator
5868+ type: object
5869+ type: array
5870+ x-kubernetes-list-type: atomic
5871+ matchLabels:
5872+ additionalProperties:
5873+ type: string
5874+ description: |-
5875+ matchLabels is a map of {key,value} pairs. A single {key,value} in the matchLabels
5876+ map is equivalent to an element of matchExpressions, whose key field is "key", the
5877+ operator is "In", and the values array contains only "value". The requirements are ANDed.
5878+ type: object
5879+ type: object
5880+ x-kubernetes-map-type: atomic
5881+ matchLabelKeys:
5882+ description: |-
5883+ MatchLabelKeys is a set of pod label keys to select the pods over which
5884+ spreading will be calculated. The keys are used to lookup values from the
5885+ incoming pod labels, those key-value labels are ANDed with labelSelector
5886+ to select the group of existing pods over which spreading will be calculated
5887+ for the incoming pod. The same key is forbidden to exist in both MatchLabelKeys and LabelSelector.
5888+ MatchLabelKeys cannot be set when LabelSelector isn't set.
5889+ Keys that don't exist in the incoming pod labels will
5890+ be ignored. A null or empty list means only match against labelSelector.
5891+
5892+ This is a beta field and requires the MatchLabelKeysInPodTopologySpread feature gate to be enabled (enabled by default).
5893+ items:
5894+ type: string
5895+ type: array
5896+ x-kubernetes-list-type: atomic
5897+ maxSkew:
5898+ description: |-
5899+ MaxSkew describes the degree to which pods may be unevenly distributed.
5900+ When `whenUnsatisfiable=DoNotSchedule`, it is the maximum permitted difference
5901+ between the number of matching pods in the target topology and the global minimum.
5902+ The global minimum is the minimum number of matching pods in an eligible domain
5903+ or zero if the number of eligible domains is less than MinDomains.
5904+ For example, in a 3-zone cluster, MaxSkew is set to 1, and pods with the same
5905+ labelSelector spread as 2/2/1:
5906+ In this case, the global minimum is 1.
5907+ | zone1 | zone2 | zone3 |
5908+ | P P | P P | P |
5909+ - if MaxSkew is 1, incoming pod can only be scheduled to zone3 to become 2/2/2;
5910+ scheduling it onto zone1(zone2) would make the ActualSkew(3-1) on zone1(zone2)
5911+ violate MaxSkew(1).
5912+ - if MaxSkew is 2, incoming pod can be scheduled onto any zone.
5913+ When `whenUnsatisfiable=ScheduleAnyway`, it is used to give higher precedence
5914+ to topologies that satisfy it.
5915+ It's a required field. Default value is 1 and 0 is not allowed.
5916+ format: int32
5917+ type: integer
5918+ minDomains:
5919+ description: |-
5920+ MinDomains indicates a minimum number of eligible domains.
5921+ When the number of eligible domains with matching topology keys is less than minDomains,
5922+ Pod Topology Spread treats "global minimum" as 0, and then the calculation of Skew is performed.
5923+ And when the number of eligible domains with matching topology keys equals or greater than minDomains,
5924+ this value has no effect on scheduling.
5925+ As a result, when the number of eligible domains is less than minDomains,
5926+ scheduler won't schedule more than maxSkew Pods to those domains.
5927+ If value is nil, the constraint behaves as if MinDomains is equal to 1.
5928+ Valid values are integers greater than 0.
5929+ When value is not nil, WhenUnsatisfiable must be DoNotSchedule.
5930+
5931+ For example, in a 3-zone cluster, MaxSkew is set to 2, MinDomains is set to 5 and pods with the same
5932+ labelSelector spread as 2/2/2:
5933+ | zone1 | zone2 | zone3 |
5934+ | P P | P P | P P |
5935+ The number of domains is less than 5(MinDomains), so "global minimum" is treated as 0.
5936+ In this situation, new pod with the same labelSelector cannot be scheduled,
5937+ because computed skew will be 3(3 - 0) if new Pod is scheduled to any of the three zones,
5938+ it will violate MaxSkew.
5939+ format: int32
5940+ type: integer
5941+ nodeAffinityPolicy:
5942+ description: |-
5943+ NodeAffinityPolicy indicates how we will treat Pod's nodeAffinity/nodeSelector
5944+ when calculating pod topology spread skew. Options are:
5945+ - Honor: only nodes matching nodeAffinity/nodeSelector are included in the calculations.
5946+ - Ignore: nodeAffinity/nodeSelector are ignored. All nodes are included in the calculations.
5947+
5948+ If this value is nil, the behavior is equivalent to the Honor policy.
5949+ This is a beta-level feature default enabled by the NodeInclusionPolicyInPodTopologySpread feature flag.
5950+ type: string
5951+ nodeTaintsPolicy:
5952+ description: |-
5953+ NodeTaintsPolicy indicates how we will treat node taints when calculating
5954+ pod topology spread skew. Options are:
5955+ - Honor: nodes without taints, along with tainted nodes for which the incoming pod
5956+ has a toleration, are included.
5957+ - Ignore: node taints are ignored. All nodes are included.
5958+
5959+ If this value is nil, the behavior is equivalent to the Ignore policy.
5960+ This is a beta-level feature default enabled by the NodeInclusionPolicyInPodTopologySpread feature flag.
5961+ type: string
5962+ topologyKey:
5963+ description: |-
5964+ TopologyKey is the key of node labels. Nodes that have a label with this key
5965+ and identical values are considered to be in the same topology.
5966+ We consider each <key, value> as a "bucket", and try to put balanced number
5967+ of pods into each bucket.
5968+ We define a domain as a particular instance of a topology.
5969+ Also, we define an eligible domain as a domain whose nodes meet the requirements of
5970+ nodeAffinityPolicy and nodeTaintsPolicy.
5971+ e.g. If TopologyKey is "kubernetes.io/hostname", each Node is a domain of that topology.
5972+ And, if TopologyKey is "topology.kubernetes.io/zone", each zone is a domain of that topology.
5973+ It's a required field.
5974+ type: string
5975+ whenUnsatisfiable:
5976+ description: |-
5977+ WhenUnsatisfiable indicates how to deal with a pod if it doesn't satisfy
5978+ the spread constraint.
5979+ - DoNotSchedule (default) tells the scheduler not to schedule it.
5980+ - ScheduleAnyway tells the scheduler to schedule the pod in any location,
5981+ but giving higher precedence to topologies that would help reduce the
5982+ skew.
5983+ A constraint is considered "Unsatisfiable" for an incoming pod
5984+ if and only if every possible node assignment for that pod would violate
5985+ "MaxSkew" on some topology.
5986+ For example, in a 3-zone cluster, MaxSkew is set to 1, and pods with the same
5987+ labelSelector spread as 3/1/1:
5988+ | zone1 | zone2 | zone3 |
5989+ | P P P | P | P |
5990+ If WhenUnsatisfiable is set to DoNotSchedule, incoming pod can only be scheduled
5991+ to zone2(zone3) to become 3/2/1(3/1/2) as ActualSkew(2-1) on zone2(zone3) satisfies
5992+ MaxSkew(1). In other words, the cluster can still be imbalanced, but scheduler
5993+ won't make it *more* imbalanced.
5994+ It's a required field.
5995+ type: string
5996+ required:
5997+ - maxSkew
5998+ - topologyKey
5999+ - whenUnsatisfiable
6000+ type: object
6001+ type: array
6002+ x-kubernetes-list-map-keys:
6003+ - topologyKey
6004+ - whenUnsatisfiable
6005+ x-kubernetes-list-type: map
58236006 updateStrategy:
58246007 description: The deployment strategy to use to replace existing
58256008 pods with new ones.
0 commit comments