Description of problem: Given a persistent volume has 'recycle' reclaim policy, and it is bound to a claim. When the claim is deleted, the persistent volume status becomes 'FAILED' immediately Version-Release number of selected component (if applicable): How reproducible: openshift v3.1.1.2 kubernetes v1.1.0-origin-1107-g4c8e6f4 etcd 2.1.2 Steps to Reproduce: 1. Create a PV with reclaim policy set as 'recycle' oc create -f https://raw.githubusercontent.com/openshift-qe/v3-testfiles/master/persistent-volumes/iscsi/pv-rwx.json 2. Create a PVC oc create -f https://raw.githubusercontent.com/openshift-qe/v3-testfiles/master/persistent-volumes/iscsi/pvc-rwx.json 3. When PV and PVC are bound, delete this PVC 4. Get PV status Actual results: After step 4: # oc get pv NAME LABELS CAPACITY ACCESSMODES STATUS CLAIM REASON AGE iscsi <none> 5Gi RWX Failed jhou/iscsic 5m # oc describe pv iscsi Name: iscsi Labels: <none> Status: Failed Claim: jhou/iscsic Reclaim Policy: Recycle Access Modes: RWX Capacity: 5Gi Message: no volume plugin matched Source: Type: ISCSI (an ISCSI Disk resource that is attached to a kubelet's host machine and then exposed to the pod) TargetPortal: 192.168.0.225:3260 IQN: iqn.2015-06.world.server:storage.target00 Lun: 0 FSType: ext4 ReadOnly: false Logs: ``` Jan 15 14:32:38 openshift-115.lab.eng.nay.redhat.com atomic-openshift-master[38864]: I0115 14:32:38.151773 38864 persistentvolume_claim_binder_controller.go:196] Synchronizing PersistentVolume[iscsi], current phase: Bound Jan 15 14:32:38 openshift-115.lab.eng.nay.redhat.com atomic-openshift-master[38864]: I0115 14:32:38.167844 38864 persistentvolume_claim_binder_controller.go:313] PersistentVolume[iscsi] changing phase from Bound to Released Jan 15 14:32:38 openshift-115.lab.eng.nay.redhat.com atomic-openshift-master[38864]: I0115 14:32:38.177292 38864 persistentvolume_recycler_controller.go:101] Reclaiming PersistentVolume[iscsi] Jan 15 14:32:38 openshift-115.lab.eng.nay.redhat.com atomic-openshift-master[38864]: I0115 14:32:38.179403 38864 persistentvolume_claim_binder_controller.go:196] Synchronizing PersistentVolume[iscsi], current phase: Released Jan 15 14:32:38 openshift-115.lab.eng.nay.redhat.com atomic-openshift-master[38864]: I0115 14:32:38.188287 38864 persistentvolume_recycler_controller.go:133] Recycling PersistentVolume[iscsi] Jan 15 14:32:38 openshift-115.lab.eng.nay.redhat.com atomic-openshift-master[38864]: I0115 14:32:38.188328 38864 persistentvolume_recycler_controller.go:165] PersistentVolume[iscsi] changing phase from Released to Failed Jan 15 14:32:38 openshift-115.lab.eng.nay.redhat.com atomic-openshift-master[38864]: I0115 14:32:38.414246 38864 persistentvolume_claim_binder_controller.go:196] Synchronizing PersistentVolume[iscsi], current phase: Failed Jan 15 14:32:38 openshift-115.lab.eng.nay.redhat.com atomic-openshift-master[38864]: I0115 14:32:38.414298 38864 persistentvolume_claim_binder_controller.go:304] PersistentVolume[iscsi] previously failed recycling. Skipping. ``` Expected results: Not sure if this is our expected behavior, since recycle policy now only supports hostpath and nfs. Additional info:
this is not a bug- iscsi has no recycler. Marking as feature request.
Yes, The ISCSI volume plugin does not support Recycle Operations on a claim release. However, I think we should throw an intuitive error & description when someone attempts to create a PV where the the PV definition specifies something that the underlying volume type does not support. Assigning to Brad Childs (Dev Lead) so that he can re-assign to a member of the dev team.
Only NFS and hostPath support recycle, any PV with recycle policy and other volume plugins will reproduce this issue.
Closing as 'wont fix'. The recycle feature is deprecated since 3.7