9.2.5 Hierarchical Snapshot Databases
Hierarchical snapshots enable you to create snapshot databases from other snapshot databases.
Oracle Exadata System Software release 12.2.1.1.0 introduced hierarchical snapshots. You might want to use hierarchical snapshots if you are working on your snapshot database and you want to save a copy before you make additional changes to it.
There is no set limit to the number of levels allowed in the hierarchy, but for performance and management reasons, a practical limit of 10 levels is recommended.
Figure 9-8 Hierarchical Snapshot Databases
![Description of Figure 9-8 follows Description of Figure 9-8 follows](img/sagug_vm_020b.png)
Description of "Figure 9-8 Hierarchical Snapshot Databases"
A snapshot database points to its parent’s blocks for data. If you make a change to a snapshot database, then the snapshot database allocates a new block for the changed data. For the unchanged data, the snapshot points to the parent’s blocks. A snapshot that is several levels descended from the original test master will retrieve its data by traversing up the tree starting with the snapshot database from which it was created.
Figure 9-9 Allocation of Blocks in Hierarchical Snapshot Databases
![Description of Figure 9-9 follows Description of Figure 9-9 follows](img/sagug_vm_021a.png)
Description of "Figure 9-9 Allocation of Blocks in Hierarchical Snapshot Databases"
If you have taken a snapshot database from another snapshot database, and you make a change to the parent snapshot database, then you have to drop all snapshot databases that depend on that snapshot database. The parent snapshot database becomes read-only when child snapshots are created from it. If you want to write to the parent snapshot again, you must drop all child snapshots.
Parent topic: Overview of Exadata Snapshots