Jobs and CronJobs
Job Example
apiVersion: batch/v1
kind: Job
metadata:
name: sum-job
spec:
parallelism: 10
completions: 2
backoffLimit: 5
template:
metadata:
name: sum-job-pod
spec:
containers:
- name: sum-job-pod-container
image: busybox
command: ["/bin/sh", "-c"]
args: ["expr 3 + 2"]
restartPolicy: OnFailure
A pod definition file named throw-dice-pod.yaml is given. The image throw-dice randomly returns a value between 1 and 6.6 is considered success and all others are failure.
Try deploying the POD and view the POD logs for the generated number.
# throw-dice-pod.yaml
apiVersion: v1
kind: Pod
metadata:
name: throw-dice-pod
spec:
containers:
- image: kodekloud/throw-dice
name: throw-dice
restartPolicy: Never
Executando Comandos… Ao observar os logs do Pod concluímos que houve uma falha já que o valor gerado(2) não foi 6, que seria considerado um sucesso. O mais estranho é que no comando describe encontramos um erro reportado pelo kubelet “MountVolume.SetUp failed for volume "kube-api-access-z6ghh" : object "default"/"kube-root-ca.crt" not registered”.
Então, o insucesso do Pod foi causado pelo valor gerado(2) ou por um problema de certificados?
root@controlplane ~ ➜ k apply -f throw-dice-pod.yaml
pod/throw-dice-pod created
root@controlplane ~ ➜ k get pods
NAME READY STATUS RESTARTS AGE
throw-dice-pod 0/1 Error 0 15s
root@controlplane ~ ➜ k logs throw-dice-pod
2
root@controlplane ~ ➜ k describe pods throw-dice-pod
Name: throw-dice-pod
Namespace: default
Status: Failed
Containers:
throw-dice:
Container ID: docker://12bf124268456067ee8035f2fe95ab...
Image: kodekloud/throw-dice
Image ID: docker-pullable://kodekloud/throw-dice@...
State: Terminated
Reason: Error
Exit Code: 1
Started: Wed, 13 Jul 2022 09:11:28 +0000
Finished: Wed, 13 Jul 2022 09:11:28 +0000
Ready: False
Restart Count: 0
Environment: <none>
Mounts:
/var/run/secrets/kubernetes.io/serviceaccount from kube-api-access-z6ghh (ro)
Conditions:
Type Status
Initialized True
Ready False
ContainersReady False
PodScheduled True
Volumes:
kube-api-access-z6ghh:
Type: Projected (a volume that contains injected data from multiple sources)
TokenExpirationSeconds: 3607
ConfigMapName: kube-root-ca.crt
ConfigMapOptional: <nil>
DownwardAPI: true
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal Scheduled 53s default-scheduler Successfully assigned default/throw-dice-pod to controlplane
Normal Pulling 51s kubelet Pulling image "kodekloud/throw-dice"
Normal Pulled 50s kubelet Successfully pulled image "kodekloud/throw-dice" in 1.552159624s
Normal Created 49s kubelet Created container throw-dice
Normal Started 47s kubelet Started container throw-dice
Warning FailedMount 42s (x3 over 44s) kubelet MountVolume.SetUp failed for volume "kube-api-access-z6ghh" : object "default"/"kube-root-ca.crt" not registered
Create a Job using this POD definition file or from the imperative command and look at how many attempts does it take to get a '6'.
Use the specification given on the below.
apiVersion: batch/v1
kind: Job
metadata:
name: throw-dice-job
spec:
parallelism: 1
completions: 1
backoffLimit: 6 # default
template:
spec:
containers:
- name: throw-dice-job-pod
image: kodekloud/throw-dice
restartPolicy: Never
root@controlplane ~ ➜ k get pods,jobs
NAME READY STATUS RESTARTS AGE
pod/throw-dice-job-9q6g2 0/1 Error 0 26s
pod/throw-dice-job-h7r8z 0/1 Error 0 30s
pod/throw-dice-job-j6ndz 0/1 Error 0 22s
pod/throw-dice-job-knznw 0/1 Completed 0 19s
pod/throw-dice-job-vz8zp 0/1 Error 0 34s
NAME COMPLETIONS DURATION AGE
job.batch/throw-dice-job 1/1 19s 34s
Detalhe, como a restartPolicy era Never são criados n Pods até que ocorra um sucesso ou o backoffLimit seja atingido.
Vamos definir o backoffLimit para 2. Foram criados dois Pods que terminaram com status de erro, neste momento o backoffLimit foi atingido e uma última tentativa de completar o Job é realizada com um terceiro e último Pod.
kubectl get pods,jobs 06:42:30
NAME READY STATUS RESTARTS AGE
pod/throw-dice-job--1-gm5wj 0/1 Error 0 31s
pod/throw-dice-job--1-k4v7c 0/1 Error 0 37s
pod/throw-dice-job--1-qvh2k 0/1 Error 0 21s
NAME COMPLETIONS DURATION AGE
job.batch/throw-dice-job 0/1 37s 37s
Update the job definition to run as many times as required to get 3 successful 6's
Delete existing job and create a new one with the given spec. Monitor and wait for the job to succeed.
apiVersion: batch/v1
kind: Job
metadata:
name: throw-dice-job
spec:
parallelism: 1
completions: 3
backoffLimit: 6 # default
template:
spec:
containers:
- name: throw-dice-job-pod
image: kodekloud/throw-dice
restartPolicy: Never
That took a while. Let us try to speed it up, by running upto 3 jobs in parallel.
Update the job definition to run 3 jobs in parallel.
apiVersion: batch/v1
kind: Job
metadata:
name: throw-dice-job
spec:
parallelism: 3
completions: 3
backoffLimit: 6 # default
template:
spec:
containers:
- name: throw-dice-job-pod
image: kodekloud/throw-dice
restartPolicy: Never
Primeira tentativa o Job atingiu o backoffLimit
root@controlplane ~ ➜ k get jobs,pods
NAME COMPLETIONS DURATION AGE
job.batch/throw-dice-job 2/3 38s 38s
NAME READY STATUS RESTARTS AGE
pod/throw-dice-job-2hr7b 0/1 Error 0 9s
pod/throw-dice-job-4vk7l 0/1 Error 0 14s
pod/throw-dice-job-5dd24 0/1 Error 0 19s
pod/throw-dice-job-h6v84 0/1 Error 0 25s
pod/throw-dice-job-kd8w8 0/1 Error 0 38s
pod/throw-dice-job-lf29l 0/1 Error 0 38s
pod/throw-dice-job-mbs9d 0/1 Completed 0 38s
pod/throw-dice-job-nz8dz 0/1 Completed 0 16s
pod/throw-dice-job-zc5vf 0/1 Error 0 22s
root@controlplane ~ ➜ k describe jobs throw-dice-job
Name: throw-dice-job
Namespace: default
Selector: controller-uid=a6d33e91-8e81-450b-8d11-0ad5a2d08537
Labels: controller-uid=a6d33e91-8e81-450b-8d11-0ad5a2d08537
job-name=throw-dice-job
Annotations: batch.kubernetes.io/job-tracking:
Parallelism: 3
Completions: 3
Completion Mode: NonIndexed
Start Time: Wed, 13 Jul 2022 09:55:11 +0000
Pods Statuses: 0 Active / 2 Succeeded / 7 Failed
Pod Template:
Labels: controller-uid=a6d33e91-8e81-450b-8d11-0ad5a2d08537
job-name=throw-dice-job
Containers:
throw-dice-job-pod:
Image: kodekloud/throw-dice
Port: <none>
Host Port: <none>
Environment: <none>
Mounts: <none>
Volumes: <none>
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal SuccessfulCreate 45s job-controller Created pod: throw-dice-job-kd8w8
Normal SuccessfulCreate 45s job-controller Created pod: throw-dice-job-mbs9d
Normal SuccessfulCreate 45s job-controller Created pod: throw-dice-job-lf29l
Normal SuccessfulCreate 32s job-controller Created pod: throw-dice-job-h6v84
Normal SuccessfulCreate 29s job-controller Created pod: throw-dice-job-zc5vf
Normal SuccessfulCreate 26s job-controller Created pod: throw-dice-job-5dd24
Normal SuccessfulCreate 23s job-controller Created pod: throw-dice-job-nz8dz
Normal SuccessfulCreate 21s job-controller Created pod: throw-dice-job-4vk7l
Normal SuccessfulCreate 16s job-controller Created pod: throw-dice-job-2hr7b
Warning BackoffLimitExceeded 12s job-controller Job has reached the specified backoff limit
Let us now schedule that job to run at 21:30 hours every day.
Create a CronJob for this
apiVersion: batch/v1
kind: CronJob
metadata:
name: throw-dice-cron-job
spec:
schedule: 30 21 * * *
jobTemplate:
spec:
parallelism: 1 # default
completions: 1 # default
backoffLimit: 6 # default
template:
spec:
containers:
- image: kodekloud/throw-dice
name: throw-dice-cron-job-container
restartPolicy: Never