资源调度

a-c-dream 发布于

labels与selector

在配置文件中可以使用labels来指定资源的labels

1
2
3
4
5
6
[root@k8s-master pods]# kubectl get po --show-labels
NAME                     READY   STATUS    RESTARTS      AGE   LABELS
nginx-85b98978db-4xdrk   1/1     Running   2 (13h ago)   19h   app=nginx,pod-template-hash=85b98978db
nginx-85b98978db-8h2hh   1/1     Running   3 (13h ago)   23h   app=nginx,pod-template-hash=85b98978db
nginx-85b98978db-h452n   1/1     Running   2 (13h ago)   19h   app=nginx,pod-template-hash=85b98978db
nginx-po                 1/1     Running   0             95s   test=1.0.0,type=app

-l可以使用selector来筛选资源，多个条件并列写时是and的关系

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
[root@k8s-master pods]# kubectl get po --show-labels
NAME                     READY   STATUS    RESTARTS      AGE   LABELS
nginx-85b98978db-4xdrk   1/1     Running   2 (13h ago)   19h   app=nginx,pod-template-hash=85b98978db
nginx-85b98978db-8h2hh   1/1     Running   3 (13h ago)   23h   app=nginx,pod-template-hash=85b98978db
nginx-85b98978db-h452n   1/1     Running   2 (13h ago)   19h   app=nginx,pod-template-hash=85b98978db
nginx-po                 1/1     Running   0             95s   test=1.0.0,type=app
[root@k8s-master pods]# vim nginx-perstop.yaml 
[root@k8s-master pods]# kubectl get po -l app=nginx
NAME                     READY   STATUS    RESTARTS      AGE
nginx-85b98978db-4xdrk   1/1     Running   2 (13h ago)   19h
nginx-85b98978db-8h2hh   1/1     Running   3 (13h ago)   24h
nginx-85b98978db-h452n   1/1     Running   2 (13h ago)   19h
[root@k8s-master pods]# kubectl get po -l test=1.0.0
NAME       READY   STATUS    RESTARTS   AGE
nginx-po   1/1     Running   0          5m11s
[root@k8s-master pods]# kubectl get po -l "test in (1.0.0,1.1.1)"
NAME       READY   STATUS    RESTARTS   AGE
nginx-po   1/1     Running   0          5m46s
[root@k8s-master pods]# kubectl get po -l test=1.1.1,type=app
No resources found in default namespace.

deployment

使用kubectl create deploy nginx-deploy --image=nginx:1.7.9创建一个deployment

查看rs和pod，发现deploy会自动创建对应的rs和pod

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
[root@k8s-master pods]# kubectl get deploy
NAME           READY   UP-TO-DATE   AVAILABLE   AGE
nginx          3/3     3            3           24h
nginx-deploy   1/1     1            1           64s
[root@k8s-master pods]# kubectl get rs
NAME                      DESIRED   CURRENT   READY   AGE
nginx-85b98978db          3         3         3       24h
nginx-deploy-78d8bf4fd7   1         1         1       3m2s
[root@k8s-master pods]# kubectl get po
NAME                            READY   STATUS    RESTARTS      AGE
nginx-85b98978db-4xdrk          1/1     Running   2 (14h ago)   20h
nginx-85b98978db-8h2hh          1/1     Running   3 (14h ago)   24h
nginx-85b98978db-h452n          1/1     Running   2 (14h ago)   20h
nginx-deploy-78d8bf4fd7-p4rkq   1/1     Running   0             4m38s
nginx-po                        1/1     Running   0             40m

使用kubectl get deploy nginx-deploy -o yaml命令可以查看他的配置文件

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
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
apiVersion: apps/v1
kind: Deployment
metadata:
  annotations:
    deployment.kubernetes.io/revision: "1"
  creationTimestamp: "2023-11-28T03:10:15Z"
  generation: 1
  labels:
    app: nginx-deploy
  name: nginx-deploy
  namespace: default
  resourceVersion: "30430"
  uid: 404bfa08-851b-4e1b-b35b-9c44131203fd
spec:
  progressDeadlineSeconds: 600
  replicas: 1
  revisionHistoryLimit: 10
  selector:
    matchLabels:
      app: nginx-deploy
  strategy:
    rollingUpdate:
      maxSurge: 25%
      maxUnavailable: 25%
    type: RollingUpdate
  template:
    metadata:
      creationTimestamp: null
      labels:
        app: nginx-deploy
    spec:
      containers:
      - image: nginx:1.7.9
        imagePullPolicy: IfNotPresent
        name: nginx
        resources: {}
        terminationMessagePath: /dev/termination-log
        terminationMessagePolicy: File
      dnsPolicy: ClusterFirst
      restartPolicy: Always
      schedulerName: default-scheduler
      securityContext: {}
      terminationGracePeriodSeconds: 30
status:
  availableReplicas: 1
  conditions:
  - lastTransitionTime: "2023-11-28T03:10:18Z"
    lastUpdateTime: "2023-11-28T03:10:18Z"
    message: Deployment has minimum availability.
    reason: MinimumReplicasAvailable
    status: "True"
    type: Available
  - lastTransitionTime: "2023-11-28T03:10:15Z"
    lastUpdateTime: "2023-11-28T03:10:18Z"
    message: ReplicaSet "nginx-deploy-78d8bf4fd7" has successfully progressed.
    reason: NewReplicaSetAvailable
    status: "True"
    type: Progressing
  observedGeneration: 1
  readyReplicas: 1
  replicas: 1
  updatedReplicas: 1

删除一些不必要的配置项之后

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
apiVersion: apps/v1 # deployment api 版本
kind: Deployment # 资源类型为 deployment
metadata: # 元信息
  labels: # 标签
    app: nginx-deploy # 具体的 key: value 配置形式
  name: nginx-deploy # deployment 的名字
  namespace: default # 所在的命名空间
spec:
  replicas: 1 # 期望副本数
  revisionHistoryLimit: 10 # 进行滚动更新后，保留的历史版本数
  selector: # 选择器，用于找到匹配的 RS
    matchLabels: # 按照标签匹配
      app: nginx-deploy # 匹配的标签key/value
  strategy: # 更新策略
    rollingUpdate: # 滚动更新配置
      maxSurge: 25% # 进行滚动更新时，更新的个数最多可以超过期望副本数的个数/比例
      maxUnavailable: 25% # 进行滚动更新时，最大不可用比例更新比例，表示在所有副本数中，最多可以有多少个不更新成功
    type: RollingUpdate # 更新类型，采用滚动更新
  template: # pod 模板
    metadata: # pod 的元信息
      labels: # pod 的标签
        app: nginx-deploy
    spec: # pod 期望信息
      containers: # pod 的容器
      - image: nginx:1.7.9 # 镜像
        imagePullPolicy: IfNotPresent # 拉取策略
        name: nginx # 容器名称
      restartPolicy: Always # 重启策略
      terminationGracePeriodSeconds: 30 # 删除操作最多宽限多长时间

滚动更新

只有修改了 deployment 配置文件中的 template 中的属性后，才会触发更新操作

使用kubectl edit deploy/nginx-deploy命令修改label属性，发现deploy并不会更新（通过kubectl get deploy查看deploy状态时，发现READY一直是1/1）

下面修改image版本号

使用describe发现deploy完成了更新操作

image-20231128140755782

把副本数修改为3

image-20231128141249039

版本回滚

有时候你可能想回退一个Deployment，例如，当Deployment不稳定时，比如一直crash looping。

ps：可以通过设置 .spec.revisonHistoryLimit 来指定 deployment 保留多少 revison，如果设置为 0，则不允许 deployment 回退了。

更新 deployment 时参数不小心写错，如 nginx:1.9.1 写成了 nginx:1.91

kubectl set image deployment/nginx-deploy nginx=nginx:1.91

此时deploy会执行更新操作，使用kubectl get po，发现刚刚更新的pod处于ErrImagePull状态（显然这是由于刚才版本号写错了导致的）

image-20231128142552409

下面进行版本回滚

通过kubectl rollout history deploy/nginx-deploy查看可以回滚的版本

image-20231128142811549

通过kubectl rollout history deploy/nginx-deploy可以查看具体的版本信息

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
40
41
42
43
[root@k8s-master ~]# kubectl rollout history deploy/nginx-deploy --revision=2
deployment.apps/nginx-deploy with revision #2
Pod Template:
  Labels:	app=nginx-deploy
	pod-template-hash=754898b577
  Containers:
   nginx:
    Image:	nginx:1.9.1
    Port:	<none>
    Host Port:	<none>
    Environment:	<none>
    Mounts:	<none>
  Volumes:	<none>

[root@k8s-master ~]# kubectl rollout history deploy/nginx-deploy --revision=1
deployment.apps/nginx-deploy with revision #1
Pod Template:
  Labels:	app=nginx-deploy
	pod-template-hash=78d8bf4fd7
  Containers:
   nginx:
    Image:	nginx:1.7.9
    Port:	<none>
    Host Port:	<none>
    Environment:	<none>
    Mounts:	<none>
  Volumes:	<none>

[root@k8s-master ~]# kubectl rollout history deploy/nginx-deploy --revision=3
deployment.apps/nginx-deploy with revision #3
Pod Template:
  Labels:	app=nginx-deploy
	pod-template-hash=f7f5656c7
  Containers:
   nginx:
    Image:	nginx:1.91
    Port:	<none>
    Host Port:	<none>
    Environment:	<none>
    Mounts:	<none>
  Volumes:	<none>

[root@k8s-master ~]#

确认要回退的版本后，可以通过kubectl rollout undo deploy/nginx-deploy --to-revision=1回退到指定版本

再次通过 kubectl get deployment 和 kubectl describe deployment 可以看到，我们的版本已经回退到对应的 revison 上了

暂停与恢复

由于每次对 pod template 中的信息发生修改后，都会触发更新 deployment 操作，那么此时如果频繁修改信息，就会产生多次更新，而实际上只需要执行最后一次更新即可，当出现此类情况时我们就可以暂停 deployment 的 rollout

通过 kubectl rollout pause deployment <name> 就可以实现暂停，直到你下次恢复后才会继续进行滚动更新

使用edit更新配置文件后 pod 和 rs均没有改变（也就是说并没有更新）

image-20231128155232654

使用kubectl rollout resume deploy nginx-deploy命令恢复

image-20231128155805347

a-c-dream

a-c-dream

贵在坚持

93
13
0
CATEGORYS
TAGS