Description of problem: In containerzied OpenShift cluster, rbd volumes could not be dynamically provisioned. PVC events logged failure reasion "rbd: create volume failed, err: executable file not found in $PATH" Version-Release number of selected component (if applicable): openshift v3.4.0.39 kubernetes v1.4.0+776c994 etcd 3.1.0-rc.0 How reproducible: Always Steps to Reproduce: 1. Setup containerized OpenShift cluster, setup ceph rbd storage server 2. Create a StorageClass for Ceph RBD provisioner apiVersion: storage.k8s.io/v1beta1 kind: StorageClass metadata: creationTimestamp: 2017-01-09T06:54:58Z name: cephrbdprovisioner resourceVersion: "8131" selfLink: /apis/storage.k8s.io/v1beta1/storageclasses/cephrbdprovisioner uid: 892a27c2-d638-11e6-800d-0eb7d2142d46 parameters: adminId: admin adminSecretName: cephrbd-secret adminSecretNamespace: default monitors: xxx:6789 pool: rbd userId: admin userSecretName: cephrbd-user-secret provisioner: kubernetes.io/rbd 3. Create a PVC to dynamically provision a PV { "kind": "PersistentVolumeClaim", "apiVersion": "v1", "metadata": { "name": "rbdc", "annotations": { "volume.beta.kubernetes.io/storage-class": "cephrbdprovisioner" } }, "spec": { "accessModes": [ "ReadWriteOnce" ], "resources": { "requests": { "storage": "5Gi" } } } } Actual results: After step 3: PVC is Pending, PV could not be provisioned. oc describe pvc rbdc Name: rbdc Namespace: jhou1 StorageClass: cephrbdprovisioner Status: Pending Volume: Labels: <none> Capacity: Access Modes: Events: FirstSeen LastSeen Count From SubobjectPath Type Reason Message --------- -------- ----- ---- ------------- -------- ------ ------- 18s 5s 2 {persistentvolume-controller } Warning ProvisioningFailed Failed to provision volume with StorageClass "cephrbdprovisioner": rbd: create volume failed, err: executable file not found in $PATH Expected results: Should provision successfully. Additional info:
This sounds like missing RBD packages on the nodes. Huamin - can you figure out the packages needed on each node?
Does this call actually nsenter to the node? Or is it looking inside the (master) container for the binaries?
@eparis: this call happens on master (provisioner), so rbd binary (in ceph-common) has to be in master image.
Verified this is fixed in openshift v3.4.1.2
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-2017:0218