NetApp NFS Driver: copy volume_to_image Uses Live Volume Copy Instead of Snapshots
Bug #2137259 reported by
Alexander Hild
This bug affects 1 person
| Affects | Status | Importance | Assigned to | Milestone | |
|---|---|---|---|---|---|
| Cinder |
New
|
High
|
Unassigned | ||
Bug Description
With the Cinder config parameter enable_
The NetApp NFS driver lacks its own copy_volume_
The default copy_volume_
This results in inconsistent data for 'in-use' volumes with active I/O, potentially leading to corrupt or incomplete images.
[1]https:/
| Changed in cinder: | |
| importance: | Undecided → High |
To post a comment you must log in.

Hi Alexander, thanks for reporting your findings.
Just confirming, have you set image_upload_ use_cinder_ backend= True in your backend in cinder.conf?
I believe this is required for the NetApp backend to handle the request as a clone operation [1] instead of going through generic nfs code.
This doc has more details: https:/ /docs.openstack .org/cinder/ latest/ admin/volume- backed- image.html# uploading- a-volume- to-image- service
[1] https:/ /opendev. org/openstack/ cinder/ src/commit/ 00505d25879992b 268402a71125ccc 65b3c4088d/ cinder/ volume/ manager. py#L1771