Queue capability can overuse

[zen-xu]

/kind bug

What happened:

I have limited the queue capability, but all jobs can be scheduled!!!

queue1 capability is 2 CPU, but can run 3 jobs each with 1CPU

What you expected to happen:

Only 2 jobs can be scheduled

How to reproduce it (as minimally and precisely as possible):

---
apiVersion: scheduling.volcano.sh/v1beta1
kind: Queue
metadata:
  name: queue1
spec:
  weight: 1
  capability:
    cpu: 2
    memory: 4Gi
---
apiVersion: batch.volcano.sh/v1alpha1
kind: Job
metadata:
  name: job1
  labels:
    volcano.sh/job-type: ElasticJob
spec:
  minAvailable: 1
  schedulerName: volcano
  queue: queue1
  tasks:
  - name: worker
    replicas: 1
    template:
      spec:
        securityContext:
          runAsUser: 1000
        terminationGracePeriodSeconds: 5
        containers:
        - image: ubuntu:16.04
          name: worker
          resources:
            requests:
              cpu: 1
              memory: 1Gi
          args: ["sleep", "1000"]
---
apiVersion: batch.volcano.sh/v1alpha1
kind: Job
metadata:
  name: job2
  labels:
    volcano.sh/job-type: ElasticJob
spec:
  minAvailable: 1
  schedulerName: volcano
  queue: queue1
  tasks:
  - name: worker
    replicas: 1
    template:
      spec:
        securityContext:
          runAsUser: 1000
        terminationGracePeriodSeconds: 5
        containers:
        - image: ubuntu:16.04
          name: worker
          resources:
            requests:
              cpu: 1
              memory: 1Gi
          args: ["sleep", "1000"]
---
apiVersion: batch.volcano.sh/v1alpha1
kind: Job
metadata:
  name: job3
  labels:
    volcano.sh/job-type: ElasticJob
spec:
  minAvailable: 1
  schedulerName: volcano
  queue: queue1
  tasks:
  - name: worker
    replicas: 1
    template:
      spec:
        securityContext:
          runAsUser: 1000
        terminationGracePeriodSeconds: 5
        containers:
        - image: ubuntu:16.04
          name: worker
          resources:
            requests:
              cpu: 1
              memory: 1Gi
          args: ["sleep", "1000"]

or replicas set 3

---
apiVersion: scheduling.volcano.sh/v1beta1
kind: Queue
metadata:
  name: queue1
spec:
  weight: 1
  capability:
    cpu: 2
    memory: 4Gi
---
apiVersion: batch.volcano.sh/v1alpha1
kind: Job
metadata:
  name: job1
  labels:
    volcano.sh/job-type: ElasticJob
spec:
  minAvailable: 1
  schedulerName: volcano
  queue: queue1
  tasks:
  - name: worker
    replicas: 3
    template:
      spec:
        securityContext:
          runAsUser: 1000
        terminationGracePeriodSeconds: 5
        containers:
        - image: ubuntu:16.04
          name: worker
          resources:
            requests:
              cpu: 1
              memory: 1Gi
          args: ["sleep", "1000"]

Anything else we need to know?:

Environment:

• Volcano Version:
  0.4
• Kubernetes version (use kubectl version):
  v1.18.2
• OS (e.g. from /etc/os-release):
  Ubuntu18.04
• Kernel (e.g. uname -a):
  Linux 4.18.0-25-generic

[hzxuzhonghu]

cc @k82cn

[stale]

Hello 👋 Looks like there was no activity on this issue for last 90 days.
Do you mind updating us on the status? Is this still reproducible or needed? If yes, just comment on this PR or push a commit. Thanks! 🤗
If there will be no activity for 60 days, this issue will be closed (we can always reopen an issue if we need!).

[hzxuzhonghu]

Duplicate with #921, fixed in v0.4.2

[zen-xu]

@hzxuzhonghu
I don't think #921 has fixed all this problem.
I have try to use v0.4.2, queue still can overuse when job specify with replicas case

---
apiVersion: scheduling.volcano.sh/v1beta1
kind: Queue
metadata:
  name: queue1
spec:
  weight: 1
  capability:
    cpu: 2
    memory: 4Gi
---
apiVersion: batch.volcano.sh/v1alpha1
kind: Job
metadata:
  name: job1
  labels:
    volcano.sh/job-type: ElasticJob
spec:
  minAvailable: 1
  schedulerName: volcano
  queue: queue1
  tasks:
  - name: worker
    replicas: 3
    template:
      spec:
        securityContext:
          runAsUser: 1000
        terminationGracePeriodSeconds: 5
        containers:
        - image: ubuntu:16.04
          name: worker
          resources:
            requests:
              cpu: 1
              memory: 1Gi
          args: ["sleep", "1000"]