(In reply to weiwei jiang from comment #0)
> Description of problem:
> Checked with cluster capacity and found it always got "Failed to set default
> scheduler config: Error in opening default scheduler config file: open : no
> such file or directory" even I give a existed path for --default-config
> 
> /bin/sh -ec /bin/cluster-capacity --default-config=/test-s/scheduler.json
> --podspec=/test-pod/pod.yaml --verbose
> 
> $ ls /test-*
> /test-pod:
> pod.yaml
> 
> /test-s:
> scheduler.json
> 
> 
> Version-Release number of selected component (if applicable):
> atomic-openshift-cluster-capacity-3.9.0-0.21.0.git.0.2a50d06.el7.x86_64
> # openshift version 
> openshift v3.9.0-0.20.0
> kubernetes v1.9.1+a0ce1bc657
> etcd 3.2.8
> 
> How reproducible:
> always
> 
> Steps to Reproduce:
> 1. Create a podspec as a configmap
> oc create configmap cluster-capacity-configmap --from-file=pod.yaml=pod.yaml
> -n default
> # cat pod.yaml 
> apiVersion: v1
> kind: Pod
> metadata:
>   creationTimestamp: null
>   name: cluster-capacity-stub-container
>   namespace: cluster-capacity
> spec:
>   containers:
>   - image: gcr.io/google_containers/pause:2.0
>     imagePullPolicy: Always
>     name: cluster-capacity-stub-container
>     resources:
>       limits:
>         cpu: 200m
>         memory: 100Mi
>       requests:
>         cpu: 100m
>         memory: 80Mi
>   dnsPolicy: Default
>   nodeSelector:
>     load: high
>     region: hpc
>   restartPolicy: OnFailure
>   schedulerName: default-scheduler
> status: {}
> 
> 2. oc create secret generic sf
> --from-file=scheduler.json=/etc/origin/master/scheduler.json  -n default

Why are you creating secret from scheduler config file? 


> 3. create a rc for cluster-capacity
> apiVersion: v1
> kind: ReplicationController
> metadata:
>   creationTimestamp: 2018-01-18T07:30:32Z
>   generation: 5
>   labels:
>     run: cluster-capacity
>   name: cluster-capacity
>   namespace: default
>   resourceVersion: "22043"
>   selfLink:
> /api/v1/namespaces/default/replicationcontrollers/cluster-capacity
>   uid: 778a86ed-fc21-11e7-80da-fa163ead188e
> spec:
>   replicas: 1
>   selector:
>     run: cluster-capacity
>   template:
>     metadata:
>       creationTimestamp: null
>       labels:
>         run: cluster-capacity
>     spec:
>       containers:
>       - command:
>         - /bin/sh
>         - -ec
>         - |
>           /bin/cluster-capacity --default-config=/test-s/scheduler.json

why are you mounting scheduler.json config file as secret?

> --podspec=/test-pod/pod.yaml --verbose;while true;do sleep 10;done
>         env:
>         - name: CC_INCLUSTER
>           value: "true"
>         image:
> brew-pulp-docker01.web.prod.ext.phx2.redhat.com:8888/openshift3/ose-cluster-
> capacity
>         imagePullPolicy: Always
>         name: cluster-capacity
>         resources: {}
>         terminationMessagePath: /dev/termination-log
>         terminationMessagePolicy: File
>         volumeMounts:
>         - mountPath: /test-pod
>           name: test-volume
>         - mountPath: /test-s
>           name: ss
>       dnsPolicy: ClusterFirst
>       restartPolicy: Always
>       schedulerName: default-scheduler
>       securityContext: {}
>       serviceAccount: cluster-capacity-sa
>       serviceAccountName: cluster-capacity-sa
>       terminationGracePeriodSeconds: 30
>       volumes:
>       - configMap:
>           defaultMode: 420
>           name: cluster-capacity-configmap
>         name: test-volume
>       - name: ss
>         secret:
>           defaultMode: 420
>           secretName: sf
> 
> 
> Actual results:
> # oc logs -n default -f cluster-capacity-bgjh5
> Failed to set default scheduler config: Error in opening default scheduler
> config file: open : no such file or directory
> 
> Expected results:
> cluster-capacity should work well
> 
> Additional info:

Also in general, you need to node provide scheduler.json unless your cluster us using some customize default scheduler.

Just want to correct my comment:
Also in general, you need NOT provide scheduler.json unless your cluster is 
using some customized default scheduler.

Also can you show me your scheduler.json? I just tested locally --default-config and it seemed to work. I will try in a pod now.

I can reproduce the issue. I am working on a fix.

PR: https://github.com/openshift/origin/pull/18198

In case you want to test it before, you could use this image: docker.io/aveshagarwal/cluster-capacity

Checked with atomic-openshift-cluster-capacity-3.9.0-0.24.0.git.0.fc8ad63.el7.x86_64, and work now.

# oc logs -f cluster-capacity-g9xfc
cluster-capacity-stub-container pod requirements:
	- CPU: 100m
	- Memory: 80Mi
	- NodeSelector: load=high,region=hpc

The cluster can schedule 0 instance(s) of the pod cluster-capacity-stub-container.

Termination reason: Unschedulable: 0/5 nodes are available: 1 NodeUnschedulable, 5 MatchNodeSelector.

According to https://bugzilla.redhat.com/show_bug.cgi?id=1535940#c6, move to verified.

Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHBA-2018:0489