LUN Mapping Not Restored After Snapshot Revert in OpenStack Cinder with NetApp ONTAP
Applies to
- NetApp ONTAP 9.16.1P10 (and similar ONTAP versions)
- OpenStack Cinder (Z release)
- Lenovo DM7200F (AFF OEM)
- Fibre Channel (FC) and iSCSI environments
- Boot/system (root) volumes in OpenStack
- Lenovo xCloud or similar OpenStack orchestration frameworks
Issue
After performing a “revert to snapshot” operation on a boot/system volume in an OpenStack-integrated environment using NetApp ONTAP as the backend, the LUN mapping is not automatically restored. This results in the volume becoming inaccessible to the original VM/host until manual intervention is performed.
Observed behavior:
- The original LUN is unmapped and destroyed during the revert process.
- A new cloned LUN is created from the snapshot.
- The new LUN is not mapped to the original host/igroup.
- All paths in multipath show as failed.
- Manual mapping is required to restore access.
- Issue is reproducible across all boot volumes; data volumes are not affected.
Relevant log output:
[ONTAP audit logs]Unmapping LUN volume-7757d269-f469-49e3-807f-fcc8902a7ad9 from the initiators ['initiator,'initiator'] terminate_connection_fc...Volume 7757d269-f469-49e3-807f-fcc8902a7ad9 reverted to snapshot successfully....After snapshot restore, the LUN is in unmapped state, but manual mapping can be performed successfully.
