在默认情况下,一个Pod在哪个Node节点上运行,是由Scheduler组件采用相应的算法计算出来的,这个过程是不受人工控制的。但是在实际使用中,这并不满足的需求,因为很多情况下,我们想控制某些Pod到达某些节点上,那么应该怎么做呢?这就要求了解kubernetes对Pod的调度规则,kubernetes提供了四大类调度方式:
spec:
affinity: # 亲和性配置
nodeAffinity: # Node亲和性
requiredDuringSchedulingIgnoredDuringExecution: # Node节点必须满足指定的所有规则才可以,相当于硬限制
nodeSelectorTerms: # 节点选择列表
matchFields: # 按节点字段列出的节点选择器要求列表
key: # 键
values: # 值
operator # 关系符 支持Exists, DoesNotExist, In, NotIn, Gt, Lt ,意义分别为包含,不包含,在···里面,不在···里面,大于,小于
matchExpressions: # 按节点标签列出的节点选择器要求列表(推荐)
key: # 键
values: # 值
operator: # 关系符 支持Exists, DoesNotExist, In, NotIn, Gt, Lt
preferredDuringSchedulingIgnoredDuringExecution: # 优先调度到满足指定的规则的Node,相当于软限制 (倾向)
preference: # 一个节点选择器项,与相应的权重相关联
matchFields: # 按节点字段列出的节点选择器要求列表
matchExpressions: #按节点标签列出的节点选择器要求列表(推荐)
key: # 键
values: # 值
operator: # 关系符 支持In, NotIn, Exists, DoesNotExist, Gt, Lt
matchFields:
weight: # 倾向权重,在范围1-100。
podAffinity: # Pod亲和性
requiredDuringSchedulingIgnoredDuringExecution: # Pod必须满足指定的所有规则才可以,相当于硬限制
namespace:
topologyKey:
labelSelector:
matchExpressions:
key:
values:
operator: # 关系符 支持Exists, DoesNotExist, In, NotIn, Gt, Lt
matchLabels:
preferredDuringSchedulingIgnoredDuringExecution: # 优先调度到满足指定的规则的Node,相当于软限制 (倾向)
podAffinityTerm:
namespace:
topologyKey:
labelSelector:
matchExpressions:
key:
values:
operator: # 关系符 支持Exists, DoesNotExist, In, NotIn, Gt, Lt
matchLabels:
weight:
podAntiAffinity: # pod反亲和性
1 topologyKey用于指定调度时作用域,例如:
如果指定为kubernetes.io/hostname,那就是以Node节点为区分范围
如果指定为beta.kubernetes.io/os,则以Node节点的操作系统类型来区分
2 关系符的使用说明:
- matchExpressions:
- key: nodeenv # 匹配存在标签的key为nodeenv的节点
operator: Exists
- key: nodeenv # 匹配标签的key为nodeenv,且value是"xxx"或"yyy"的节点
operator: In
values: ["xxx","yyy"]
- key: nodeenv # 匹配标签的key为nodeenv,且value大于"xxx"的节点
operator: Gt
values: "xxx"
定向调度,指的是利用在pod上声明nodeName或者nodeSelector,以此将Pod调度到期望的node节点上。注意,这里的调度是强制的,这就意味着即使要调度的目标Node不存在,也会向上面进行调度,只不过pod运行失败而已。
NodeName用于强制约束将Pod调度到指定的Name的Node节点上。这种方式,其实是直接跳过Scheduler的调度逻辑,直接将Pod调度到指定名称的节点。
# 创建一个pod-nodename.yaml文件
apiVersion: v1
kind: Pod
metadata:
name: pod-nodename
namespace: dev
spec:
containers:
- name: nginx
image: nginx:1.17.1
nodeName: node1 # 指定调度到node1节点上
# 创建Pod
[root@k8s-master01 ~]# kubectl create -f pod-nodename.yaml
pod/pod-nodename created
# 查看Pod调度到NODE属性,确实是调度到了node1节点上
[root@k8s-master01 ~]# kubectl get pods pod-nodename -n dev -o wide
NAME READY STATUS RESTARTS AGE IP NODE ......
pod-nodename 1/1 Running 0 56s 10.244.1.87 node1 ......
# 接下来,删除pod,修改nodeName的值为node3(并没有node3节点)
[root@k8s-master01 ~]# kubectl delete -f pod-nodename.yaml
pod "pod-nodename" deleted
[root@k8s-master01 ~]# vim pod-nodename.yaml
[root@k8s-master01 ~]# kubectl create -f pod-nodename.yaml
pod/pod-nodename created
# 再次查看,发现已经向Node3节点调度,但是由于不存在node3节点,所以pod无法正常运行
[root@k8s-master01 ~]# kubectl get pods pod-nodename -n dev -o wide
NAME READY STATUS RESTARTS AGE IP NODE ......
pod-nodename 0/1 Pending 0 6s node3 ......
NodeSelector用于将pod调度到添加了指定标签的node节点上。它是通过kubernetes的label-selector机制实现的,也就是说,在pod创建之前,会由scheduler使用MatchNodeSelector调度策略进行label匹配,找出目标node,然后将pod调度到目标节点,该匹配规则是强制约束。
# 首先分别为node节点添加标签
[root@k8s-master01 ~]# kubectl label nodes node1 nodeenv=pro
node/node2 labeled
[root@k8s-master01 ~]# kubectl label nodes node2 nodeenv=test
node/node2 labeled
# 创建一个pod-nodeselector.yaml文件,并使用它创建Pod
apiVersion: v1
kind: Pod
metadata:
name: pod-nodeselector
namespace: dev
spec:
containers:
- name: nginx
image: nginx:1.17.1
nodeSelector:
nodeenv: pro # 指定调度到具有nodeenv=pro标签的节点上
# 创建Pod
[root@k8s-master01 ~]# kubectl create -f pod-nodeselector.yaml
pod/pod-nodeselector created
# 查看Pod调度到NODE属性,确实是调度到了node1节点上
[root@k8s-master01 ~]# kubectl get pods pod-nodeselector -n dev -o wide
NAME READY STATUS RESTARTS AGE IP NODE ......
pod-nodeselector 1/1 Running 0 47s 10.244.1.87 node1 ......
# 接下来,删除pod,修改nodeSelector的值为nodeenv: xxxx(不存在打有此标签的节点)
[root@k8s-master01 ~]# kubectl delete -f pod-nodeselector.yaml
pod "pod-nodeselector" deleted
[root@k8s-master01 ~]# vim pod-nodeselector.yaml
[root@k8s-master01 ~]# kubectl create -f pod-nodeselector.yaml
pod/pod-nodeselector created
# 再次查看,发现pod无法正常运行,Node的值为none
[root@k8s-master01 ~]# kubectl get pods -n dev -o wide
NAME READY STATUS RESTARTS AGE IP NODE
pod-nodeselector 0/1 Pending 0 2m20s
# 查看详情,发现node selector匹配失败的提示
[root@k8s-master01 ~]# kubectl describe pods pod-nodeselector -n dev
.......
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Warning FailedScheduling default-scheduler 0/3 nodes are available: 3 node(s) didn't match node selector.
定向调度的方式,使用起来非常方便,但是也有一定的问题,那就是如果没有满足条件的Node,那么Pod将不会被运行,即使在集群中还有可用Node列表也不行,这就限制了它的使用场景。
基于上面的问题,kubernetes还提供了一种亲和性调度(Affinity)。它在NodeSelector的基础之上的进行了扩展,可以通过配置的形式,实现优先选择满足条件的Node进行调度,如果没有,也可以调度到不满足条件的节点上,使调度更加灵活。
Affinity主要分为三类:
关于亲和性(反亲和性)使用场景的说明:
首先来看一下NodeAffinity的可配置项:
spec:
affinity: # 亲和性配置
nodeAffinity: # Node亲和性
requiredDuringSchedulingIgnoredDuringExecution: # Node节点必须满足指定的所有规则才可以,相当于硬限制
nodeSelectorTerms: # 节点选择列表
matchFields: # 按节点字段列出的节点选择器要求列表
key: # 键
values: # 值
operator # 关系符 支持Exists, DoesNotExist, In, NotIn, Gt, Lt ,意义分别为包含,不包含,在···里面,不在···里面,大于,小于
matchExpressions: # 按节点标签列出的节点选择器要求列表(推荐)
key: # 键
values: # 值
operator: # 关系符 支持Exists, DoesNotExist, In, NotIn, Gt, Lt
preferredDuringSchedulingIgnoredDuringExecution: # 优先调度到满足指定的规则的Node,相当于软限制 (倾向)
preference: # 一个节点选择器项,与相应的权重相关联
matchFields: # 按节点字段列出的节点选择器要求列表
matchExpressions: #按节点标签列出的节点选择器要求列表(推荐)
key: # 键
values: # 值
operator: # 关系符 支持In, NotIn, Exists, DoesNotExist, Gt, Lt
matchFields:
weight: # 倾向权重,在范围1-100。
关系符的使用说明:
- matchExpressions:
- key: nodeenv # 匹配存在标签的key为nodeenv的节点
operator: Exists
- key: nodeenv # 匹配标签的key为nodeenv,且value是"xxx"或"yyy"的节点
operator: In
values: ["xxx","yyy"]
- key: nodeenv # 匹配标签的key为nodeenv,且value大于"xxx"的节点
operator: Gt
values: "xxx"
接下来首先演示一下requiredDuringSchedulingIgnoredDuringExecution
# 创建pod-nodeaffinity-required.yaml
apiVersion: v1
kind: Pod
metadata:
name: pod-nodeaffinity-required
namespace: dev
spec:
containers:
- name: nginx
image: nginx:1.17.1
affinity: #亲和性设置
nodeAffinity: #设置node亲和性
requiredDuringSchedulingIgnoredDuringExecution: # 硬限制
nodeSelectorTerms:
- matchExpressions: # 匹配env的值在["xxx","yyy"]中的标签
- key: nodeenv
operator: In
values: ["xxx","yyy"]
# 创建pod
[root@k8s-master01 ~]# kubectl create -f pod-nodeaffinity-required.yaml
pod/pod-nodeaffinity-required created
# 查看pod状态 (运行失败)
[root@k8s-master01 ~]# kubectl get pods pod-nodeaffinity-required -n dev -o wide
NAME READY STATUS RESTARTS AGE IP NODE ......
pod-nodeaffinity-required 0/1 Pending 0 16s ......
# 查看Pod的详情
# 发现调度失败,提示node选择失败
[root@k8s-master01 ~]# kubectl describe pod pod-nodeaffinity-required -n dev
......
Warning FailedScheduling default-scheduler 0/3 nodes are available: 3 node(s) didn't match node selector.
Warning FailedScheduling default-scheduler 0/3 nodes are available: 3 node(s) didn't match node selector.
#接下来,停止pod
[root@k8s-master01 ~]# kubectl delete -f pod-nodeaffinity-required.yaml
pod "pod-nodeaffinity-required" deleted
# 修改文件,将values: ["xxx","yyy"]------> ["pro","yyy"]
[root@k8s-master01 ~]# vim pod-nodeaffinity-required.yaml
# 再次启动
[root@k8s-master01 ~]# kubectl create -f pod-nodeaffinity-required.yaml
pod/pod-nodeaffinity-required created
# 此时查看,发现调度成功,已经将pod调度到了node1上
[root@k8s-master01 ~]# kubectl get pods pod-nodeaffinity-required -n dev -o wide
NAME READY STATUS RESTARTS AGE IP NODE ......
pod-nodeaffinity-required 1/1 Running 0 11s 10.244.1.89 node1 ......
接下来再演示一下requiredDuringSchedulingIgnoredDuringExecution
# 创建pod-nodeaffinity-preferred.yaml
apiVersion: v1
kind: Pod
metadata:
name: pod-nodeaffinity-preferred
namespace: dev
spec:
containers:
- name: nginx
image: nginx:1.17.1
affinity: #亲和性设置
nodeAffinity: #设置node亲和性
preferredDuringSchedulingIgnoredDuringExecution: # 软限制
- weight: 1
preference:
matchExpressions: # 匹配env的值在["xxx","yyy"]中的标签(当前环境没有)
- key: nodeenv
operator: In
values: ["xxx","yyy"]
# 创建pod
[root@k8s-master01 ~]# kubectl create -f pod-nodeaffinity-preferred.yaml
pod/pod-nodeaffinity-preferred created
# 查看pod状态 (运行成功)
[root@k8s-master01 ~]# kubectl get pod pod-nodeaffinity-preferred -n dev
NAME READY STATUS RESTARTS AGE
pod-nodeaffinity-preferred 1/1 Running 0 40s
NodeAffinity规则设置的注意事项:
- 如果同时定义了nodeSelector和nodeAffinity,那么必须两个条件都得到满足,Pod才能运行在指定的Node上
- 如果nodeAffinity指定了多个nodeSelectorTerms,那么只需要其中一个能够匹配成功即可
- 如果一个nodeSelectorTerms中有多个matchExpressions ,则一个节点必须满足所有的才能匹配成功
- 如果一个pod所在的Node在Pod运行期间其标签发生了改变,不再符合该Pod的节点亲和性需求,则系统将忽略此变化
PodAffinity主要实现以运行的Pod为参照,实现让新创建的Pod跟参照pod在一个区域的功能。
首先来看一下PodAffinity的可配置项:
spec:
affinity: # 亲和性配置
podAffinity: # Pod亲和性
requiredDuringSchedulingIgnoredDuringExecution: # Pod必须满足指定的所有规则才可以,相当于硬限制
namespace:
topologyKey:
labelSelector:
matchExpressions:
key:
values:
operator: # 关系符 支持Exists, DoesNotExist, In, NotIn, Gt, Lt
matchLabels:
preferredDuringSchedulingIgnoredDuringExecution: # 优先调度到满足指定的规则的Node,相当于软限制 (倾向)
podAffinityTerm:
namespace:
topologyKey:
labelSelector:
matchExpressions:
key:
values:
operator: # 关系符 支持Exists, DoesNotExist, In, NotIn, Gt, Lt
matchLabels:
weight:
podAntiAffinity: # pod反亲和性
# topologyKey用于指定调度时作用域,例如:
如果指定为kubernetes.io/hostname,那就是以Node节点为区分范围
如果指定为beta.kubernetes.io/os,则以Node节点的操作系统类型来区分
接下来,演示下requiredDuringSchedulingIgnoredDuringExecution,
# (1)首先创建一个参照Pod,pod-podaffinity-target.yaml:
apiVersion: v1
kind: Pod
metadata:
name: pod-podaffinity-target
namespace: dev
labels:
podenv: pro #设置标签
spec:
containers:
- name: nginx
image: nginx:1.17.1
nodeName: node1 # 将目标pod名确指定到node1上
# 启动目标pod
[root@k8s-master01 ~]# kubectl create -f pod-podaffinity-target.yaml
pod/pod-podaffinity-target created
# 查看pod状况
[root@k8s-master01 ~]# kubectl get pods pod-podaffinity-target -n dev
NAME READY STATUS RESTARTS AGE
pod-podaffinity-target 1/1 Running 0 4s
# (2)创建pod-podaffinity-required.yaml,内容如下:
apiVersion: v1
kind: Pod
metadata:
name: pod-podaffinity-required
namespace: dev
spec:
containers:
- name: nginx
image: nginx:1.17.1
affinity: #亲和性设置
podAffinity: #设置pod亲和性
requiredDuringSchedulingIgnoredDuringExecution: # 硬限制
- labelSelector:
matchExpressions: # 匹配env的值在["xxx","yyy"]中的标签
- key: podenv
operator: In
values: ["xxx","yyy"]
topologyKey: kubernetes.io/hostname
#上面配置表达的意思是:新Pod必须要与拥有标签nodeenv=xxx或者nodeenv=yyy的pod在同一Node上,显然现在没有这样pod,接下来,运行测试一下。
# 启动pod
[root@k8s-master01 ~]# kubectl create -f pod-podaffinity-required.yaml
pod/pod-podaffinity-required created
# 查看pod状态,发现未运行
[root@k8s-master01 ~]# kubectl get pods pod-podaffinity-required -n dev
NAME READY STATUS RESTARTS AGE
pod-podaffinity-required 0/1 Pending 0 9s
# 查看详细信息
[root@k8s-master01 ~]# kubectl describe pods pod-podaffinity-required -n dev
......
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Warning FailedScheduling default-scheduler 0/3 nodes are available: 2 node(s) didn't match pod affinity rules, 1 node(s) had taints that the pod didn't tolerate.
# 接下来修改 values: ["xxx","yyy"]----->values:["pro","yyy"]
# 意思是:新Pod必须要与拥有标签nodeenv=xxx或者nodeenv=yyy的pod在同一Node上
[root@k8s-master01 ~]# vim pod-podaffinity-required.yaml
# 然后重新创建pod,查看效果
[root@k8s-master01 ~]# kubectl delete -f pod-podaffinity-required.yaml
pod "pod-podaffinity-required" deleted
[root@k8s-master01 ~]# kubectl create -f pod-podaffinity-required.yaml
pod/pod-podaffinity-required created
# 发现此时Pod运行正常
[root@k8s-master01 ~]# kubectl get pods pod-podaffinity-required -n dev
NAME READY STATUS RESTARTS AGE LABELS
pod-podaffinity-required 1/1 Running 0 6s
PodAntiAffinity主要实现以运行的Pod为参照,让新创建的Pod跟参照pod不在一个区域中的功能。
它的配置方式和选项跟PodAffinty是一样的,这里不再做详细解释。
#(1)继续使用上个案例中目标pod
[root@k8s-master01 ~]# kubectl get pods -n dev -o wide --show-labels
NAME READY STATUS RESTARTS AGE IP NODE LABELS
pod-podaffinity-required 1/1 Running 0 3m29s 10.244.1.38 node1
pod-podaffinity-target 1/1 Running 0 9m25s 10.244.1.37 node1 podenv=pro
#(2)创建pod-podantiaffinity-required.yaml,内容如下:
apiVersion: v1
kind: Pod
metadata:
name: pod-podantiaffinity-required
namespace: dev
spec:
containers:
- name: nginx
image: nginx:1.17.1
affinity: #亲和性设置
podAntiAffinity: #设置pod亲和性
requiredDuringSchedulingIgnoredDuringExecution: # 硬限制
- labelSelector:
matchExpressions: # 匹配podenv的值在["pro"]中的标签
- key: podenv
operator: In
values: ["pro"]
topologyKey: kubernetes.io/hostname
#上面配置表达的意思是:新Pod必须要与拥有标签nodeenv=pro的pod不在同一Node上,运行测试一下。
# 创建pod
[root@k8s-master01 ~]# kubectl create -f pod-podantiaffinity-required.yaml
pod/pod-podantiaffinity-required created
# 查看pod
# 发现调度到了node2上
[root@k8s-master01 ~]# kubectl get pods pod-podantiaffinity-required -n dev -o wide
NAME READY STATUS RESTARTS AGE IP NODE ..
pod-podantiaffinity-required 1/1 Running 0 30s 10.244.1.96 node2 ..
调度方式都是站在Pod的角度上,通过在Pod上添加属性,来确定Pod是否要调度到指定的Node上,其实我们也可以站在Node的角度上,通过在Node上添加污点属性,来决定是否允许Pod调度过来。
Node被设置上污点之后就和Pod之间存在了一种相斥的关系,进而拒绝Pod调度进来,甚至可以将已经存在的Pod驱逐出去。
污点的格式为:key=value:effect, key和value是污点的标签,effect描述污点的作用,支持如下三个选项:
使用kubectl设置和去除污点的命令示例如下:
# 设置污点
kubectl taint nodes node1 key=value:effect
# 去除污点
kubectl taint nodes node1 key:effect-
# 去除所有污点
kubectl taint nodes node1 key-
接下来,演示下污点的效果:
# 为node1设置污点(PreferNoSchedule)
[root@k8s-master01 ~]# kubectl taint nodes node1 tag=heima:PreferNoSchedule
# 创建pod1
[root@k8s-master01 ~]# kubectl run taint1 --image=nginx:1.17.1 -n dev
[root@k8s-master01 ~]# kubectl get pods -n dev -o wide
NAME READY STATUS RESTARTS AGE IP NODE
taint1-7665f7fd85-574h4 1/1 Running 0 2m24s 10.244.1.59 node1
# 为node1设置污点(取消PreferNoSchedule,设置NoSchedule)
[root@k8s-master01 ~]# kubectl taint nodes node1 tag:PreferNoSchedule-
[root@k8s-master01 ~]# kubectl taint nodes node1 tag=heima:NoSchedule
# 创建pod2
[root@k8s-master01 ~]# kubectl run taint2 --image=nginx:1.17.1 -n dev
[root@k8s-master01 ~]# kubectl get pods taint2 -n dev -o wide
NAME READY STATUS RESTARTS AGE IP NODE
taint1-7665f7fd85-574h4 1/1 Running 0 2m24s 10.244.1.59 node1
taint2-544694789-6zmlf 0/1 Pending 0 21s
# 为node1设置污点(取消NoSchedule,设置NoExecute)
[root@k8s-master01 ~]# kubectl taint nodes node1 tag:NoSchedule-
[root@k8s-master01 ~]# kubectl taint nodes node1 tag=heima:NoExecute
# 创建pod3
[root@k8s-master01 ~]# kubectl run taint3 --image=nginx:1.17.1 -n dev
[root@k8s-master01 ~]# kubectl get pods -n dev -o wide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED
taint1-7665f7fd85-htkmp 0/1 Pending 0 35s
taint2-544694789-bn7wb 0/1 Pending 0 35s
taint3-6d78dbd749-tktkq 0/1 Pending 0 6s
我们可以在node上添加污点用于拒绝pod调度上来,但是如果就是想将一个pod调度到一个有污点的node上去,这时候应该怎么做呢?这就要使用到容忍。
污点就是拒绝,容忍就是忽略,Node通过污点拒绝pod调度上去,Pod通过容忍忽略拒绝
下面看一下容忍的详细配置:
[root@k8s-master01 ~]# kubectl explain pod.spec.tolerations
......
FIELDS:
key # 对应着要容忍的污点的键,空意味着匹配所有的键
value # 对应着要容忍的污点的值
operator # key-value的运算符,支持Equal和Exists(默认)
effect # 对应污点的effect,空意味着匹配所有影响
tolerationSeconds # 容忍时间, 当effect为NoExecute时生效,表示pod在Node上的停留时间
下面先通过一个案例看下效果:
前面已经在node1节点上打上了NoExecute的污点,此时pod是调度不上去的,可以通过给pod添加容忍,然后将其调度上去
# 创建pod-toleration.yaml,内容如下
apiVersion: v1
kind: Pod
metadata:
name: pod-toleration
namespace: dev
spec:
containers:
- name: nginx
image: nginx:1.17.1
tolerations: # 添加容忍
- key: "tag" # 要容忍的污点的key
operator: "Equal" # 操作符
value: "heima" # 容忍的污点的value
effect: "NoExecute" # 添加容忍的规则,这里必须和标记的污点规则相同
# 添加容忍之前的pod
[root@k8s-master01 ~]# kubectl get pods -n dev -o wide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED
pod-toleration 0/1 Pending 0 3s
# 添加容忍之后的pod
[root@k8s-master01 ~]# kubectl get pods -n dev -o wide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED
pod-toleration 1/1 Running 0 3s 10.244.1.62 node1