| Restoring only part of the Protection Group |
Database files are at different points in time and recovery may fail. |
Restore or clone every database volume from the same Protection Group snapshot. |
| Including OCR/voting volumes |
Cluster instability, node eviction, or unnecessary clusterware exposure. |
Exclude GRID, CRS, and +OCR volumes from database snapshot and restore procedures. |
| Presenting clone volumes to source hosts |
ASM identity conflicts or accidental production exposure. |
Use target-only host groups and verify membership before presentation. |
| Mounting source and clone ASM disk groups on one cluster |
Duplicate ASM disk group identity and possible incorrect volume selection. |
Use an alternate cluster, or ensure only one complete storage set is discoverable and active. |
| Assuming snapshot type does not affect recovery |
Incorrect recovery, failed open, or loss of required redo. |
Record whether the source snapshot was application-consistent or crash-consistent and follow the approved recovery plan. |
| Using OPEN RESETLOGS by default |
New incarnation, backup implications, and avoidable recovery complexity. |
Use it only when the selected Oracle recovery procedure requires it. |
| Device permissions or udev mismatch on target |
ASM cannot discover clone disks. |
Validate udev, ASMLib, ASMFD, multipath, and ownership rules on every RAC node before mounting. |
| Same database identity in a clone |
Clusterware, monitoring, backup, and application conflicts. |
Set a unique DB_UNIQUE_NAME; evaluate DBID, services, password file, and all integrations. |
| Assuming array volume names equal ASM disk names |
Incorrect volume mapping. |
Verify each device using WWID for SCSI or NGUID for NVMe-oF and retain an auditable mapping. |