Job
Job负责批量处理短暂的一次性任务 (short lived one-off tasks)，即仅执行一次的任务，它保证批处理任务的一个或多个Pod成功结束。
Kubernetes支持以下几种Job：
- 非并行Job：通常创建一个Pod直至其成功结束
- 固定结束次数的Job：设置.spec.completions，创建多个Pod，直到.spec.completions个Pod成功结束
- 带有工作队列的并行Job：设置.spec.Parallelism但不设置.spec.completions，当所有Pod结束并且至少一个成功时，Job就认为是成功
根据.spec.completions和.spec.Parallelism的设置，可以将Job划分为以下几种pattern：
| Job类型 | 使用示例 | 行为 | completions | Parallelism |
|---|---|---|---|---|
| 一次性Job | 数据库迁移 | 创建一个Pod直至其成功结束 | 1 | 1 |
| 固定结束次数的Job | 处理工作队列的Pod | 依次创建一个Pod运行直至completions个成功结束 | 2+ | 1 |
| 固定结束次数的并行Job | 多个Pod同时处理工作队列 | 依次创建多个Pod运行直至completions个成功结束 | 2+ | 2+ |
| 并行Job | 多个Pod同时处理工作队列 | 创建一个或多个Pod直至有一个成功结束 | 1 | 2+ |
Job Controller
Job Controller负责根据Job Spec创建Pod，并持续监控Pod的状态，直至其成功结束。如果失败，则根据restartPolicy（只支持OnFailure和Never，不支持Always）决定是否创建新的Pod再次重试任务。
Job Spec格式
- spec.template格式同Pod
- RestartPolicy仅支持Never或OnFailure
- 单个Pod时，默认Pod成功运行后Job即结束
- .spec.completions标志Job结束需要成功运行的Pod个数，默认为1
- .spec.parallelism标志并行运行的Pod的个数，默认为1
- spec.activeDeadlineSeconds标志失败Pod的重试最大时间，超过这个时间不会继续重试
一个简单的例子：
[root@k8s-master mnt]# cat job.yaml
apiVersion: batch/v1
kind: Job
metadata:
name: pi
spec:
template:
metadata:
name: pi
spec:
containers:
- name: pi
image: perl
command: ["perl", "-Mbignum=bpi", "-wle", "print bpi(2000)"]
restartPolicy: Never
[root@k8s-master mnt]#
[root@k8s-master mnt]# kubectl create -f job.yaml job.batch/pi created [root@k8s-master mnt]# kubectl get pod NAME READY STATUS RESTARTS AGE deamonset-example-vdzjp 1/1 Running 0 19m deamonset-example-xxt2z 1/1 Running 0 20m pi-7s4vx 0/1 ContainerCreating 0 5s [root@k8s-master mnt]# kubectl get pod NAME READY STATUS RESTARTS AGE deamonset-example-vdzjp 1/1 Running 0 19m deamonset-example-xxt2z 1/1 Running 0 20m pi-7s4vx 0/1 ContainerCreating 0 7s [root@k8s-master mnt]# kubectl describe pod pi-7s4vx Name: pi-7s4vx Namespace: default Priority: 0 Node: k8s-node02/192.168.180.136 Start Time: Mon, 23 Dec 2019 20:21:25 +0800 Labels: controller-uid=51af6eb5-9166-49eb-9328-8f5c6bcfe18f job-name=pi Annotations: <none> Status: Pending IP: IPs: <none> Controlled By: Job/pi Containers: pi: Container ID: Image: perl Image ID: Port: <none> Host Port: <none> Command: perl -Mbignum=bpi -wle print bpi(2000) State: Waiting Reason: ContainerCreating Ready: False Restart Count: 0 Environment: <none> Mounts: /var/run/secrets/kubernetes.io/serviceaccount from default-token-6wcrh (ro) Conditions: Type Status Initialized True Ready False ContainersReady False PodScheduled True Volumes: default-token-6wcrh: Type: Secret (a volume populated by a Secret) SecretName: default-token-6wcrh Optional: false QoS Class: BestEffort Node-Selectors: <none> Tolerations: node.kubernetes.io/not-ready:NoExecute for 300s node.kubernetes.io/unreachable:NoExecute for 300s Events: Type Reason Age From Message ---- ------ ---- ---- ------- Normal Scheduled <unknown> default-scheduler Successfully assigned default/pi-7s4vx to k8s-node02 Normal Pulling 19s kubelet, k8s-node02 Pulling image "perl" [root@k8s-master mnt]# kubectl get job NAME COMPLETIONS DURATION AGE pi 0/1 26s 26s [root@k8s-master mnt]# kubectl get job NAME COMPLETIONS DURATION AGE pi 0/1 31s 31s [root@k8s-master mnt]# kubectl describe pod pi-7s4vx Name: pi-7s4vx Namespace: default Priority: 0 Node: k8s-node02/192.168.180.136 Start Time: Mon, 23 Dec 2019 20:21:25 +0800 Labels: controller-uid=51af6eb5-9166-49eb-9328-8f5c6bcfe18f job-name=pi Annotations: <none> Status: Pending IP: IPs: <none> Controlled By: Job/pi Containers: pi: Container ID: Image: perl Image ID: Port: <none> Host Port: <none> Command: perl -Mbignum=bpi -wle print bpi(2000) State: Waiting Reason: ContainerCreating Ready: False Restart Count: 0 Environment: <none> Mounts: /var/run/secrets/kubernetes.io/serviceaccount from default-token-6wcrh (ro) Conditions: Type Status Initialized True Ready False ContainersReady False PodScheduled True Volumes: default-token-6wcrh: Type: Secret (a volume populated by a Secret) SecretName: default-token-6wcrh Optional: false QoS Class: BestEffort Node-Selectors: <none> Tolerations: node.kubernetes.io/not-ready:NoExecute for 300s node.kubernetes.io/unreachable:NoExecute for 300s Events: Type Reason Age From Message ---- ------ ---- ---- ------- Normal Scheduled <unknown> default-scheduler Successfully assigned default/pi-7s4vx to k8s-node02 Normal Pulling 2m11s kubelet, k8s-node02 Pulling image "perl"
[root@k8s-master mnt]# kubectl get pod NAME READY STATUS RESTARTS AGE deamonset-example-vdzjp 1/1 Running 0 28m deamonset-example-xxt2z 1/1 Running 0 29m pi-7s4vx 0/1 Completed 0 9m7s [root@k8s-master mnt]# kubectl get pod -o wide NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES deamonset-example-vdzjp 1/1 Running 0 28m 10.244.2.22 k8s-node01 <none> <none> deamonset-example-xxt2z 1/1 Running 0 29m 10.244.1.23 k8s-node02 <none> <none> pi-7s4vx 0/1 Completed 0 9m15s 10.244.1.24 k8s-node02 <none> <none> [root@k8s-master mnt]# kubectl logs pi-7s4vx 3.141592653589793238462643383279502884197169399375105820974944592307816678316527120190914564856692346034
固定结束次数的Job示例
apiVersion: batch/v1
kind: Job
metadata:
name: busybox
spec:
completions: 3
template:
metadata:
name: busybox
spec:
containers:
- name: busybox
image: busybox
command: ["echo", "hello"]
restartPolicy: Never
CronJob
CronJob即定时任务，就类似于Linux系统的crontab，在指定的时间周期运行指定的任务。在Kubernetes 1.5，使用CronJob需要开启batch/v2alpha1 API，即–runtime-config=batch/v2alpha1。
CronJob Spec
- .spec.schedule指定任务运行周期，格式同Cron
- .spec.jobTemplate指定需要运行的任务，格式同Job
- .spec.startingDeadlineSeconds指定任务开始的截止期限
- .spec.concurrencyPolicy指定任务的并发策略，支持Allow、Forbid和Replace三个选项
[root@k8s-master mnt]# cat cronjob.yaml
apiVersion: batch/v1beta1
kind: CronJob
metadata:
name: hello
spec:
schedule: "*/1 * * * *"
jobTemplate:
spec:
template:
spec:
containers:
- name: hello
image: busybox
args:
- /bin/sh
- -c
- date; echo Hello from the Kubernetes cluster
restartPolicy: OnFailure
[root@k8s-master mnt]#
[root@k8s-master mnt]# vim cronjob.yaml [root@k8s-master mnt]# kubectl create -f cronjob.yaml cronjob.batch/hello created [root@k8s-master mnt]# kubectl get cronjob NAME SCHEDULE SUSPEND ACTIVE LAST SCHEDULE AGE hello */1 * * * * False 0 <none> 16s [root@k8s-master mnt]# kubectl get job NAME COMPLETIONS DURATION AGE hello-1577105220 0/1 2s 2s pi 1/1 5m23s 25m [root@k8s-master mnt]# kubectl get pod NAME READY STATUS RESTARTS AGE deamonset-example-vdzjp 1/1 Running 0 45m deamonset-example-xxt2z 1/1 Running 0 47m hello-1577105220-9vjjn 0/1 Completed 0 50s pi-7s4vx 0/1 Completed 0 26m [root@k8s-master mnt]# kubectl logs pod hello-1577105220-9vjjn Error from server (NotFound): pods "pod" not found [root@k8s-master mnt]# kubectl logs hello-1577105220-9vjjn Mon Dec 23 12:47:14 UTC 2019 Hello from the Kubernetes cluster [root@k8s-master mnt]# kubectl get job -w NAME COMPLETIONS DURATION AGE hello-1577105220 1/1 9s 2m4s hello-1577105280 1/1 8s 64s hello-1577105340 0/1 4s 4s pi 1/1 5m23s 27m hello-1577105340 1/1 6s 6s hello-1577105400 0/1 0s hello-1577105400 0/1 0s 0s hello-1577105400 1/1 5s 5s ^Z [1]+ 已停止 kubectl get job -w [root@k8s-master mnt]# kubectl delete -f cronjob.yaml cronjob.batch "hello" deleted [root@k8s-master mnt]# kubectl delete -f job.yaml job.batch "pi" deleted [root@k8s-master mnt]#