mc admin heal

The mc admin heal command scans for objects that are damaged or corrupted and heals those objects.

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 AIStor:

• 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.
• aggressively heals objects after drive replacement.

Refer to Object Healing for more details on how MinIO AIStor heals objects.

mc admin heal myaistor/mybucket/myprefix

mc admin heal [FLAGS] TARGET             \
                      [--all-drives, -a] \
                      [--cross-pool]     \
                      [--dry-run, -n]    \
                      [--force]          \
                      [--pool int]       \
                      [--recursive, -r]  \
                      [--remove]         \
                      [--scan "string"]  \
                      [--set int]        \
                      [--verbose, -v]

Required

The full path to the bucket or bucket prefix on which the command should perform object healing. Specify the alias of a configured MinIO AIStor deployment as the prefix for the path. For example:

mc admin heal myaistor/mybucket/myprefix

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

Alias: -a

Optional

Select all drives and show verbose information.

Optional

Combines objects that have become split across pools.

Specifying this flag disables all other healing.

Alias: -n

Optional

Inspects the data without mutating it.

Optional

Disables warning prompts.

Optional

Heals only the given pool, specified as an integer. Pools are numbered from 1.

Alias: -r

Optional

Heals recursively.

Optional

Removes dangling objects encountered during the heal sequence.

Optional

Selects the healing scan mode. Specify one of normal, deep, or uncommitted.

The uncommitted mode heals one erasure set at a time. Pass both --pool and --set when you specify this mode.

Optional

Heals only the given erasure set, specified as an integer. Sets are numbered from 1 within each pool.

Pass --pool alongside this flag to identify a single erasure set. Used on its own, --set heals the set with that number in every pool.

Alias: -v

Optional

Show detailed information including:

• Per-server and per-drive status with tolerance information.
• Drive offline trackers showing offline event history with timestamps, duration, reason, and decision for each event.
• Server failure tolerance summary per pool.

If multiple drives within the same erasure set require healing, MinIO AIStor prioritizes the first drive that begins healing within the erasure set. Other drives in the erasure set wait until healing finishes on the first drive before the next drive starts healing.

This allows drives that begin healing to complete the process as quickly as possible.

The default output displays two sections:

Healing Status

Shows drives that are actively healing or queued for healing, organized by pool. For each drive, the output shows:

• Hostname and drive path.
• State (HEALING or QUEUED).
• Healing reason (for example, FreshDisk).
• Progress percentage (for actively healing drives).
• Set number.
• Time since healing started.

The default output shows up to 5 drives. Use --verbose for the full list.

Summary

Shows aggregate healing statistics:

• Objects healed (count and size)
• Objects failed
• Heal rate (objects per second and throughput)
• Offline disk count (if any)
• Sets exceeding parity tolerance (if any)