AdvancedCronJob支持展示上一次执行的成功job数

[qiankunli]

acj 常规输出

➜  ~ k get acj -o wide
NAME        SCHEDULE    TYPE           LASTSCHEDULETIME   AGE
clean-log   0 0 * * *   BroadcastJob   4h4m               3d20h

我们发现部分node 可能是因为负载较高的缘故，从未执行过job（pod 状态为OutOfpods），但一直没发现。

期待可以输出最近一次执行的 成功数/node节点数，来判断在所有节点上是否都执行，如果未执行，可以尽快去采取一些措施。该数据也建议与kube-state-metrics 集成，方便做监控。

[FillZpp]

@qiankunli AdvancedCronJob 只负责周期性创建 BroadcastJob 或 Job，因此输出只包含周期信息。至于每次 BroadcastJob 执行的结果，可以通过 k get bcj -o wide 来看到。

[qiankunli]

@qiankunli AdvancedCronJob 只负责周期性创建 BroadcastJob 或 Job，因此输出只包含周期信息。至于每次 BroadcastJob 执行的结果，可以通过 k get bcj -o wide 来看到。

➜  ~ k get bcj -o wide
NAME                   DESIRED   ACTIVE   SUCCEEDED   FAILED   AGE
clean-log-1611273600   52        0        10          2        5h39m

DESIRED=52，SUCCEEDED=10，FAILED=2 中间的差值在哪看呢？一般有啥原因不？

[FillZpp]

@qiankunli YAML please. k get bcj clean-log-1611273600 -o yaml

[qiankunli]

k get bcj clean-log-1611273600 -o yaml

apiVersion: apps.kruise.io/v1alpha1
kind: BroadcastJob
metadata:
  annotations:
    apps.kruise.io/scheduled-at: "2021-01-22T00:00:00Z"
  creationTimestamp: "2021-01-22T00:00:00Z"
  generation: 1
  managedFields:
  - apiVersion: apps.kruise.io/v1alpha1
    fieldsType: FieldsV1
    fieldsV1:
      f:metadata:
        f:annotations:
          .: {}
          f:apps.kruise.io/scheduled-at: {}
        f:ownerReferences: {}
      f:spec:
        .: {}
        f:completionPolicy:
          .: {}
          f:type: {}
        f:failurePolicy: {}
        f:template:
          .: {}
          f:metadata:
            .: {}
            f:creationTimestamp: {}
            f:labels:
              .: {}
              f:broadcastjob-controller-uid: {}
              f:broadcastjob-name: {}
          f:spec:
            .: {}
            f:containers: {}
            f:dnsPolicy: {}
            f:restartPolicy: {}
            f:schedulerName: {}
            f:securityContext: {}
            f:terminationGracePeriodSeconds: {}
            f:volumes: {}
      f:status:
        .: {}
        f:active: {}
        f:completionTime: {}
        f:conditions: {}
        f:desired: {}
        f:failed: {}
        f:phase: {}
        f:startTime: {}
        f:succeeded: {}
    manager: manager
    operation: Update
    time: "2021-01-22T00:01:17Z"
  name: clean-log-1611273600
  namespace: default
  ownerReferences:
  - apiVersion: apps.kruise.io/v1alpha1
    blockOwnerDeletion: true
    controller: true
    kind: AdvancedCronJob
    name: clean-log
    uid: 557e4e1a-e842-488a-bf5b-aa708d7c0c71
  resourceVersion: "215856730"
  selfLink: /apis/apps.kruise.io/v1alpha1/namespaces/default/broadcastjobs/clean-log-1611273600
  uid: 0d7551e0-8ddb-42a5-8266-229d85949cbb
spec:
  completionPolicy:
    type: Always
  failurePolicy:
    type: FailFast
  parallelism: 2147483647
  template:
    metadata:
      creationTimestamp: null
    spec:
      containers:
      - command:
        - /auto-del-1-days-ago.sh
        image: xx/clean-log:20210119
        imagePullPolicy: Always
        name: clean-log
        resources: {}
        terminationMessagePath: /dev/termination-log
        terminationMessagePolicy: File
        volumeMounts:
        - mountPath: /var/log/app
          name: log
      dnsPolicy: ClusterFirst
      restartPolicy: Never
      schedulerName: default-scheduler
      securityContext: {}
      terminationGracePeriodSeconds: 30
      volumes:
      - hostPath:
          path: /var/log/app
          type: ""
        name: log
status:
  active: 0
  completionTime: "2021-01-22T00:01:17Z"
  conditions:
  - lastProbeTime: "2021-01-22T00:01:17Z"
    lastTransitionTime: "2021-01-22T00:01:17Z"
    message: failure policy is FailurePolicyTypeFailFast and failed pod is found
    reason: Failed
    status: "True"
    type: Failed
  desired: 52
  failed: 2
  phase: failed
  startTime: "2021-01-22T00:00:01Z"
  succeeded: 10

这个acj的作用是，每天早上8点执行clean-log image 中包含的clean-log shell 清理物理机xx 天前的日志。我发现有一些pod 报Outpod（估计是物理机负载较大导致pod 创建失败），导致pod failed，如果 FailurePolicyType默认为FailFast，导致整个bj 执行失败？

[FillZpp]

@qiankunli 是的，默认 FailFast 发现有 job 失败整个 BroadcastJob 就算作失败了，你可以在 broadcastJobTemplate 里把这个策略显式设置为 Continue

[qiankunli]

@qiankunli 是的，默认 FailFast 发现有 job 失败整个 BroadcastJob 就算作失败了，你可以在 broadcastJobTemplate 里把这个策略显式设置为 Continue

@FillZpp 感谢提醒，我试试

[FillZpp]

@qiankunli 方便的话可以登记一下使用 #289 ，以便我们后续收集反馈