Tutorial: Building with CleaveDB

Drop & Restore Buckets

In addition to managing individual records, CleaveDB provides bucket-level lifecycle commands. When decommissioning a service, archiving historical datasets, or migrating between environments, you can take entire buckets offline and restore them when needed.

Soft drop an entire bucket

To take a bucket and all its contained documents offline:

DROP BUCKET staging_logs

A dropped bucket is marked as inactive. Subsequent reads and writes targeting the bucket will return an error, but the underlying data remains safely stored in the engine for disaster recovery.

Restore a dropped bucket

To bring a dropped bucket back online with all its original records intact:

RESTORE BUCKET staging_logs

CleaveDB re-activates the bucket partition, re-attaches its B+Tree indexes, and restores normal read and write availability immediately.

Permanently incinerating a bucket

To permanently obliterate an entire bucket and free all associated disk storage:

INCINERATE BUCKET deprecated_v1

This operation irreversibly removes the bucket schema, validation guards, all member documents, vector indices, and graph pointers.