k8s nodeSelector&affinity

1.分配pod到node的方法
通过node label selector实现约束pod运行到指定节点,有两种方法 nodeSelector 以及affinity 
2.nodeSelector 是k8s早起提供的节点选择器实现
1）首先为nodes打对应的label
kubectl label nodes master disktype=ssd
2)创建yaml文件

cat >> nginx-pod.yaml << EOFapiVersion: v1kind: Podmetadata:  name: nginx  labels:    env: testspec:  containers:  - name: nginx    image: nginxEOF

创建pod
kubectl create -f nginx-pod.yaml
节点默认的label
kubectl get nodes -o yaml 可以看到默认节点的label,也可以使用这些label进行约束

alpha.kubernetes.io/fluentd-ds-ready: "true"beta.kubernetes.io/arch: amd64beta.kubernetes.io/os: linuxdisktype: ssdkubeadm.alpha.kubernetes.io/role: masterkubernetes.io/hostname: master

3, affinity 
根据限制方式有两种affinity：
requiredDuringSchedulingIgnoredDuringExecution （硬限制）
条件必须满足从而调度到一个节点
preferredDuringSchedulingIgnoredDuringExecution （软限制）
条件不满足则调度到不满足条件的节点

IgnoredDuringExecution意味着当节点label改变后，pod将继续运行在这个不满足labelselector的node上面，将来会提供RequiredDuringExecution，让node驱逐不满足affinity条件的pod

根据类型有

NodeAffinity
PodAffinity
PodAntiAffinity

NodeAffinity

下面这个例子将运行pod到kubernetes.io/e2e-az-name值为e2e-az1 或e2e-az2的节点上面

apiVersion: v1kind: Podmetadata:  name: with-node-affinity  annotations:    scheduler.alpha.kubernetes.io/affinity: >      {        "nodeAffinity": {          "requiredDuringSchedulingIgnoredDuringExecution": {            "nodeSelectorTerms": [              {                "matchExpressions": [                  {                    "key": "kubernetes.io/e2e-az-name",                    "operator": "In",                    "values": ["e2e-az1", "e2e-az2"]                  }                ]              }            ]          }        }      }    another-annotation-key: another-annotation-valuespec:  containers:  - name: with-node-affinity    image: gcr.io/google_containers/pause:2.0

我们可以看到NodeSelectorTerms可以有多个，之间是或的关系，满足任意一个既满足，，MatchExpressions也可以有多个，他们之间是且的关系 必须都满足

preferredDuringSchedulingIgnoredDuringExecution 例子：apiVersion: v1kind: Podmetadata:  name: with-labels  annotations:    scheduler.alpha.kubernetes.io/affinity: >    {       "nodeAffinity": {          "preferredDuringSchedulingIgnoredDuringExecution": [                {                    "weight" : 1,                    "preference": {                        "matchExpressions": [                            {                                "key": "disktype",                                "operator": "In",                                "values": ["ssd"]                            }                        ]                    }                }            ]        }    }    another-annotation-key: another-annotation-valuespec:  containers:  - name: test-affinity    env:      - name: MYSQL_ROOT_PASSWORD        value: pass    image: mysql

preferredDuringSchedulingIgnoredDuringExecution值为列表，根据权重决定顺序
MatchExpressions 值为列表 关系为且，必须都满足

Inter-pod affinity and anti-affinity
Inter-pod affinity 是根据通过已运行在节点上的pod的标签而不是node的标签来决定被调度pod的运行节点，因为pod运行在指定的namespace所以需要自己指定运行pod的namesapce
anti-affinity 和Inter-pod affinity相反

这个例子意思是首先node需要有 failure-domain.beta.kubernetes.io/zone label 不能有kubernetes.io/hostname label  切改节点上需要有一个pod 有一个security=S1的label

apiVersion: v1kind: Podmetadata:  name: with-pod-affinity  annotations:    scheduler.alpha.kubernetes.io/affinity: >        {          "podAffinity": {            "requiredDuringSchedulingIgnoredDuringExecution": [              {                "labelSelector": {                  "matchExpressions": [                    {                      "key": "security",                      "operator": "In",                      "values": ["S1"]                    }                  ]                },                "topologyKey": "failure-domain.beta.kubernetes.io/zone"             }            ]           },          "podAntiAffinity": {            "requiredDuringSchedulingIgnoredDuringExecution": [              {                "labelSelector": {                  "matchExpressions": [                    {                      "key": "security",                      "operator": "In",                      "values": ["S2"]                    }                  ]                },                "topologyKey": "kubernetes.io/hostname"             }            ]           }         }spec:  containers:  - name: with-pod-affinity    image: gcr.io/google_containers/pause:2.0

具体字段的定义以及关系可以参考k8s源码的定义

type Affinity struct {    NodeAffinity *NodeAffinity    PodAffinity *PodAffinity    PodAntiAffinity *PodAntiAffinity}type NodeAffinity struct {    RequiredDuringSchedulingIgnoredDuringExecution *NodeSelector    PreferredDuringSchedulingIgnoredDuringExecution []PreferredSchedulingTerm}type PodAffinity struct {    RequiredDuringSchedulingIgnoredDuringExecution []PodAffinityTerm    PreferredDuringSchedulingIgnoredDuringExecution []WeightedPodAffinityTerm}type PodAntiAffinity struct {    RequiredDuringSchedulingIgnoredDuringExecution []PodAffinityTerm    PreferredDuringSchedulingIgnoredDuringExecution []WeightedPodAffinityTerm}type PreferredSchedulingTerm struct {    Weight int32    Preference NodeSelectorTerm}type NodeSelector struct {    NodeSelectorTerms []NodeSelectorTerm}type NodeSelectorTerm struct {    MatchExpressions []NodeSelectorRequirement}type NodeSelectorRequirement struct {    Key string    Operator NodeSelectorOperator    Values []string}const (    NodeSelectorOpIn           NodeSelectorOperator = "In"    NodeSelectorOpNotIn        NodeSelectorOperator = "NotIn"    NodeSelectorOpExists       NodeSelectorOperator = "Exists"    NodeSelectorOpDoesNotExist NodeSelectorOperator = "DoesNotExist"    NodeSelectorOpGt           NodeSelectorOperator = "Gt"    NodeSelectorOpLt           NodeSelectorOperator = "Lt")type PodAffinityTerm struct {    LabelSelector *metav1.LabelSelector    Namespaces []string    TopologyKey string}type LabelSelector struct {    MatchLabels map[string]string     MatchExpressions []LabelSelectorRequirement }type LabelSelectorRequirement struct {    Key string     Operator LabelSelectorOperator     Values []string }const (    LabelSelectorOpIn           LabelSelectorOperator = "In"    LabelSelectorOpNotIn        LabelSelectorOperator = "NotIn"    LabelSelectorOpExists       LabelSelectorOperator = "Exists"    LabelSelectorOpDoesNotExist LabelSelectorOperator = "DoesNotExist")type WeightedPodAffinityTerm struct {    Weight int32    PodAffinityTerm PodAffinityTerm}type PodAffinityTerm struct {    LabelSelector *metav1.LabelSelector    Namespaces []string    TopologyKey string}