# mc admin heal

LLMS index: [llms.txt](/llms.txt)

---

<a id="mc-admin-heal"></a>

<a id="command-mc.admin.heal"></a>

## Description {#description}

The [`mc admin heal`](#command-mc.admin.heal) command scans for objects that are damaged or corrupted and heals those objects.

[`mc admin heal`](#command-mc.admin.heal) is resource intensive and typically not required as a manual process, even after drive failures or corruption events.

As a part of normal operations, MinIO:

- automatically heals objects damaged by silent bit rot corruption, drive failure, or other issues on each `POST` or `GET` operation.
- performs periodic background object healing using the [scanner](/operations/concepts/scanner/#minio-concepts-scanner).
- aggressively heals objects after drive replacement.

Refer to [Object Healing](/operations/concepts/healing/#minio-concepts-healing) for more details on how MinIO heals objects.

> [!NOTE]
> **Server compatibility**
>
> The supported server for [`mcli admin`](/reference/minio-mc-admin/#command-mc.admin) is SILO. Compatibility with unmodified upstream MinIO is best effort; general S3 compatibility alone does not provide these administrative APIs.

## Syntax {#syntax}

[`mc admin heal`](#command-mc.admin.heal) has the following syntax:

```shell
mc admin heal [FLAGS] TARGET             \
                      [--all-drives, -a] \
                      [--dry-run, -n]    \
                      [--force]          \
                      [--verbose, -v]
```

[`mc admin heal`](#command-mc.admin.heal) supports the following arguments:

#### `TARGET` {#mc.admin.heal.TARGET}

*mc-cmd*

*Required*

The full path to the bucket or bucket prefix on which the command should perform object healing. Specify the [`alias`](/reference/minio-mc/mc-alias/#command-mc.alias) of a configured MinIO deployment as the prefix for the path. For example:

```shell
mc admin heal play/mybucket/myprefix
```

If the `TARGET` bucket or bucket prefix has an active healing scan, the command returns the status of that scan.

#### `--all-drives, -a` {#mc.admin.heal.-all-drives}

*mc-cmd*

*Optional*

Select all drives and show verbose information.

#### `--dry-run, -n` {#mc.admin.heal.-dry-run}

*mc-cmd*

*Optional*

Inspects objects and reports drive and shard health without modifying or deleting data on disk.

#### `--force` {#mc.admin.heal.-force}

*mc-cmd*

*Optional*

Disables warning prompts.

#### `--verbose, -v` {#mc.admin.heal.-verbose}

*mc-cmd*

*Optional*

Show information about offline and faulty healing drives.

## Behavior {#behavior}

In [SILO 20261109](/blog/release/silo-20261109/) (`RELEASE.2026-11-09T00-00-00Z`) and [mcli 20261109](/blog/release/mcli-20261109/):

- `mc admin heal --dry-run` (`-n`) remains strictly read-only and does not delete dangling or corrupted data on disk ([PR #280](https://github.com/pgsty/silo/pull/280)).
- The server classifies `xl.meta` checksum failures for healing so damaged metadata shards are repaired ([#309](https://github.com/pgsty/silo/issues/309)).
- Object versions whose queued lifecycle expiry is refused due to incomplete or damaged history are queued for healing ([#303](https://github.com/pgsty/silo/issues/303)).
- `mcli` formats heal byte progress cleanly through `EiB`.

<a id="minio-concepts-healing-colors"></a>

## Healing Colors {#healing-colors}

Some versions of MinIO used a color key as a way to differentiate objects with different healing statuses.

> [!NOTE]
> **Changed: mc**
>
> RELEASE.2024-11-17T19-35-25Z

The color meaning has been updated.

- Green indicates the bucket is healthy.
- Yellow indicates the bucket requires healing on one or more drives.
- Red indicates one or more drives are unhealthy.
- Grey indicates an indeterminate healing state.