警告
本文最后更新于 2020-12-29 11:57，文中内容可能已过时。
node 亲和性 (nodeAffinity)：选择调度到相同或不同的 node 节点
pod 亲和性 (podAffinity)：和某些 pod 调度到同一节点
pod 反亲和性 (podAntiAffinity)：和某些 pod 调度在不同节点
指定了将 pod 调度到一个节点上必须满足的规则，不满足则会处于 pending 状态，一直进行重试，直到满足为止。
优先满足指定的规则，如果不满足则会调度到其他节点上。如果有多个软策略还可以设置优先级 (weight)
使用节点亲和性硬策略将 pod 调度到 kubernetes.io/hostname 标签不是 test-k8s-node2 的节点上，软策略将 pod 尽量调度到 test-k8s-node3 节点上
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
| kind: Deployment
apiVersion: apps/v1
metadata:
name: affinity
labels:
app: nginx
spec:
replicas: 4
selector:
matchLabels:
app: nginx-affinity
template:
metadata:
labels:
app: nginx-affinity
spec:
containers:
- name: nginx
image: nginx:1.14.2
ports:
- name: web
containerPort: 80
affinity:
nodeAffinity:
requiredDuringSchedulingIgnoredDuringExecution:
nodeSelectorTerms:
- matchExpressions:
- key: kubernetes.io/hostname
operator: NotIn
values:
- test-k8s-node2
preferredDuringSchedulingIgnoredDuringExecution:
- weight: 1
preference:
matchExpressions:
- key: kubernetes.io/hostname
operator: In
values:
- test-k8s-node3
|
> operator 可选操作：`In`,`NotIn`,`Exists`,`DoesNotExist`,`Gt`,`Lt`
我这里的区域标记是：
1
2
3
4
| kubectl label nodes test-k8s-node1 zone=1
kubectl label nodes test-k8s-node3 zone=1
kubectl label nodes test-k8s-node2 zone=2
kubectl label nodes test-k8s-node4 zone=2
|
首先找到 pod 中包含 app=nginx 节点，查看 zone 标签的值 (topologyKey 的作用)，假如 zone=2，那么该 pod 会部署在 zone=2 的区域中。
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
| kind: Deployment
apiVersion: apps/v1
metadata:
name: affinity
labels:
app: nginx
spec:
replicas: 4
selector:
matchLabels:
app: nginx-affinity
template:
metadata:
labels:
app: nginx-affinity
spec:
containers:
- name: nginx
image: nginx:1.14.2
ports:
- name: web
containerPort: 80
affinity:
podAffinity:
requiredDuringSchedulingIgnoredDuringExecution:
- labelSelector:
matchExpressions:
- key: app
operator: In
values:
- nginx
topologyKey: zone
|
直接部署此文件：
均处于 pending 状态，
创建一个标签为 app=nginx 的 pod
1
2
3
4
5
6
7
8
9
10
11
12
13
| apiVersion: v1
kind: Pod
metadata:
name: test-nginx-pod
labels:
app: nginx
spec:
containers:
- name: nginx
image: nginx:1.14.2
ports:
- name: http
containerPort: 80
|
可以看到这个 pod 调度到了 test-k8s-node1 节点上，node1 节点的 zone 值为 1，所以刚才的 pod 会调度到 zone=1 的节点上
当前 app=nginx 的 pod 被调度在 node4 节点上
创建反亲和性的 pod
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
| kind: Deployment
apiVersion: apps/v1
metadata:
name: affinity
labels:
app: nginx
spec:
replicas: 4
selector:
matchLabels:
app: nginx-affinity
template:
metadata:
labels:
app: nginx-affinity
spec:
containers:
- name: nginx
image: nginx:1.14.2
ports:
- name: web
containerPort: 80
affinity:
podAntiAffinity:
requiredDuringSchedulingIgnoredDuringExecution:
- topologyKey: kubernetes.io/hostname
labelSelector:
matchExpressions:
- key: app
operator: In
values:
- nginx
|
可以看到所有 pod 都不会被调度到 node4 节点上了
SoulChild
收录于 kubernetes
微信号
微信打赏