Note: This bug is displayed in read-only format because the product is no longer active in Red Hat Bugzilla.

Bug 2039274

Summary:	rbd: snapshot can't be deleted/unmanaged if its source volume is deleted in ceph
Product:	Red Hat OpenStack	Reporter:	Takashi Kajinami <tkajinam>
Component:	openstack-cinder	Assignee:	Eric Harney <eharney>
Status:	CLOSED ERRATA	QA Contact:	Yosi Ben Shimon <ybenshim>
Severity:	low	Docs Contact:	RHOS Documentation Team <rhos-docs>
Priority:	low
Version:	16.1 (Train)	CC:	abishop, eharney, jschluet, ltoscano, mariel, pgrist, vkoul
Target Milestone:	z3	Keywords:	Triaged
Target Release:	17.1
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:
Fixed In Version:	openstack-cinder-18.2.2-17.1.20231011140828.el9ost	Doc Type:	If docs needed, set a value
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2024-05-22 20:39:13 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:
Embargoed:
Bug Depends On:
Bug Blocks:	2160510

Description Takashi Kajinami 2022-01-11 12:05:10 UTC

Description of problem:

A user encountered the situation where they need to clear data in their ceph cluster and started the cluster from scratch.
After re-initializing the cluster, they tried to remove a snapshot record left in cinder db but both delete and unmanage operation keep failing.

It seems the problem is that the rbd driver tries to access the source volume first, and the operations fails if the source volume doesn't exist.


Version-Release number of selected component (if applicable):
RHOSP16.1.6

How reproducible:
Always

Steps to Reproduce:
1. Create a volume and a snapshot
2. Delete rbd images in the backend ceph pool
3. Delete/Unmanage the snapshot

Actual results:
The snapshot becomes error_deleting/error_unmanaging status

Expected results:
The snapshot is deleted/unmanaged

Additional info:

Comment 2 Takashi Kajinami 2022-01-12 03:23:37 UTC

We solved the issue by creating an empty dummy image in backend.

 # rbd create --size 1 <pool>/volume-<volume id>

We also observe that deleting a volume gets stuck in deleting status
when the actual image doesn't exist in rbd. I'll check some details
and open a separate bug if needed.

Comment 3 Alan Bishop 2022-01-12 21:01:12 UTC

There's a root cause for this, so no need for a separate bug (per comment #2).

The RBD driver does have provision for gracefully handling a request to delete a volume when it's already deleted on the backend, see [1]. The driver also has similar code to handle deleting a snapshot when the snapshot no long exists [2]. However, what's missing is code to handle an error when deleting a snapshot because the parent volume doesn't exist. We need to catch failures that occur at [3].

[1] https://opendev.org/openstack/cinder/src/branch/master/cinder/volume/drivers/rbd.py#L1192
[2] https://opendev.org/openstack/cinder/src/branch/master/cinder/volume/drivers/rbd.py#L1315
[3] https://opendev.org/openstack/cinder/src/branch/master/cinder/volume/drivers/rbd.py#L1289

BTW, deleting the volume fails because of the occur that occurs when first it tries to delete the volume's snapshots.

Comment 4 Rajat Dhasmana 2022-09-23 03:44:36 UTC

*** Bug 2117852 has been marked as a duplicate of this bug. ***

Comment 16 errata-xmlrpc 2024-05-22 20:39:13 UTC

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 (Red Hat OpenStack Platform 17.1.3 bug fix and enhancement 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-2024:2741