While reviewing the Databricks official documentation and performing a POC on managed volumes, I observed that volumes cannot be recovered using the UNDROP command if accidentally deleted — unlike managed tables.

Technically, this recovery should work for volumes as well, and here’s why:

1. Retention period is the same for both tables and volumes (default 7 days).

2. Data and metadata for both are stored in the underlying storage during the retention period and are automatically purged only after the retention expires.

If Databricks is retaining volume data for 7 days post-deletion, it logically implies that recovery should be possible — just like tables. If recovery is not supported, then the concept of retention for volumes after deletion becomes redundant.

This is a gap in the current behavior and Databricks should ensure consistency between tables and volumes when it comes to retention and recovery mechanisms.

Thanks.