Bug 1296035

Summary:	Recycle policy for ceph backend (RBD recycler)
Product:	OpenShift Container Platform	Reporter:	Miheer Salunke <misalunk>
Component:	Storage	Assignee:	Bradley Childs <bchilds>
Status:	CLOSED NOTABUG	QA Contact:	Liang Xia <lxia>
Severity:	low	Docs Contact:
Priority:	low
Version:	3.1.0	CC:	aos-bugs, bchilds, jkrieger, knakayam
Target Milestone:	---	Keywords:	Reopened
Target Release:	---
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2016-11-17 21:51:33 UTC	Type:	Bug
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:

Description Miheer Salunke 2016-01-06 09:05:48 UTC

Description of problem:

We've seen OSE3 documentation talks about recycling volumes is only supported in NFS and HostPath ( https://docs.openshift.org/latest/architecture/additional_concepts/storage.html#pv-recycling-policy ) but Ceph backend shows recycle as a persistentVolumeReclaimPolicy.  A bugzilla ticket is opened to fix documentation: https://bugzilla.redhat.com/show_bug.cgi?id=1295460

But, the thing is, what will be the procedure to manually recycle the volume? Should the admin user delete the pv, then delete the volume in the storage backend and recreate it?

The recycling not implemented when using ceph backend. AFAIK, the recycling is just a special pod (ose-recycler) that uses the previously released volume and runs an "rm -Rf /" isn't it?

I guess we can also have generic recycler that works with all volumes type or in this case it will be better to have rbd recycler.

If we set the reclaimpolicy to recycle for ceph pv, logs say:

messages:Jan  4 04:56:16 ose1master atomic-openshift-master: I0104 04:56:16.019862    9525 persistentvolume_recycler_controller.go:94] Reclaiming PersistentVolume[ceph-vol13]
messages:Jan  4 04:56:16 ose1master atomic-openshift-master: I0104 04:56:16.023230    9525 persistentvolume_recycler_controller.go:126] Recycling PersistentVolume[ceph-vol13]
messages:Jan  4 04:56:16 ose1master atomic-openshift-master: E0104 04:56:16.023259    9525 persistentvolume_recycler_controller.go:118] Could not recycle volume spec: Could not find recyclable volume plugin for spec: no volume plugin matched


Version-Release number of selected component (if applicable):
3.1

How reproducible:
Always

Steps to Reproduce:
1. As per https://docs.openshift.org/latest/architecture/additional_concepts/storage.html#pv-recycling-policy
2.
3.

Actual results:
Setting the reclaimpolicy to recycle for ceph pv fails.

Expected results:
Setting the reclaimpolicy to recycle for ceph pv shall be successful.

Additional info:

Comment 1 Mark Turansky 2016-01-11 15:27:20 UTC

This is a feature request for a ceph recycler.

Huamin Chen has created a prototype for a Ceph provisioner, which would create and delete volumes on demand.  A recycler is probably not required.

Comment 11 Steve Watt 2016-11-17 21:51:33 UTC

The Ceph RBD and Cinder Dynamic Provisioners are part of the OpenShift 3.4 release and will not provide the the ability to recycle volumes as the recycling functionality will no longer be needed because dynamically provisioned volumes are deleted when the claim is released.