Safe Administrative Storage Operations

Use this page before destructive or bulk work in Ceph Admin or Storage Ops.

Before you act

• Confirm the workspace and endpoint/context shown in the topbar.
• Review the exact target list after filters and selections are applied.
• Keep destructive RGW options such as purge-data, purge-objects, or bypass-gc disabled unless the operation explicitly requires them.
• Read the confirmation dialog and type the exact requested phrase only after its target and effect match your intent.

Administrative operations

Operation	Main risk	Safety expectation
Purge bucket contents	Permanently removes current objects, versions, and delete markers.	Explicit target summary, confirmation phrase, progress and failures.
Delete RGW account or user	Removes an administrative identity; optional data purge can remove owned data.	Unitary target confirmation and persistent RGW result.
Delete an RGW bucket	Can remove all bucket data when purge options are enabled.	Destructive options off by default and exact target confirmation.
Link or unlink a bucket	Changes bucket ownership association.	Review old and new owner; do not treat it as an object ACL repair.
Check or fix bucket index	fix modifies RGW bucket index state.	Bulk checks stay read-only; unitary repair requires explicit confirmation.
Bulk lifecycle/notification changes	Changes configuration on many buckets.	Preview/apply workflow and visible per-target result.

Treat the RGW HTTP status and Ceph error code shown by Ceph Admin as the result of the administrative request. A backend request can be accepted while RGW still returns a domain failure such as BucketNotEmpty or AccountNotEmpty.

After the operation

Review completed and failed targets, refresh the affected administrative list, and verify the resulting state before retrying any failure. Use the application audit for the control-plane action and backend logs when deeper diagnosis is required.

Related pages

• Bucket purge
• Ceph Admin
• Storage Ops
• Troubleshooting

Visual example