Object Management

An object is binary data, such as images, audio files, spreadsheets, or even binary executable code. The term “Binary Large Object” or “blob” is sometimes associated to object storage, although blobs can be anywhere from a few bytes to several terabytes in size. Object Storage platforms like MinIO provide dedicated tools and capabilities for storing, listing, and retrieving objects using a standard S3-compatible API.

MinIO Object Storage uses buckets to organize objects. A bucket is similar to a top-level drive, folder, or directory in a filesystem (/mnt/data or C:\), where each bucket can hold an arbitrary number of objects.

The structure of objects on the MinIO server might look similar to the following:

/ #root

With the example structure, an administrator would create the /images, /videos and /articles buckets. Client applications write objects to those buckets using the full “path” to that object, including all intermediate prefixes.

MinIO supports multiple levels of nested directories and objects using prefixes to support even the most dynamic object storage workloads. MinIO automatically infers the intermediate prefixes, such as /articles/john.doe from the full object path using / as a delimiter. Clients and administrators should not create these prefixes manually.

Neither clients nor administrators would manually create the intermediate prefixes, as MinIO automatically infers them from the object name.

Path vs Virtual Host Bucket Access

MinIO supports both path-style (default) or virtual-host bucket lookups.

For example, consider a MinIO deployment with an assigned Fully Qualified Domain Name (FQDN) of

  • With path-style lookups, applications specify the full path to a bucket, such as

  • With virtual-host lookups, applications specify the bucket as a subdomain, such as

Some applications may require or expect virtual-host lookup support when performing S3 operations against MinIO. To enable virtual-host bucket lookup, you must set the MINIO_DOMAIN environment variable to a FQDN that resolves to the MinIO Deployment.

If you configure MINIO_DOMAIN, you must consider all subdomains of the specified FQDN as exclusively assigned for use as bucket names. Any MinIO services which conflict with those domains, such as replication targets, may exhibit unexpected or undesired behavior as a result of the collision.

For example, if setting, you cannot assign any subdomains of (in the form of * to any MinIO service or target. This includes hostnames for use with bucket, batch, or site replication.


For deployments with TLS enabled, you must ensure your TLS certificate SANs cover all subdomains of the leftmost domain specified to MINIO_DOMAIN.

For example, the example of requires a TLS SAN that covers the subdomains of You can set an additional TLS SAN of * to appropriately cover the subdomain namespace.

TLS Wildcard rules prevent chaining to additional subdomain levels, such that a TLS certificate with a wildcard SAN of * would not cover the virtual host lookups at *

Object Organization and Planning

Administrators typically control the creation and configuration of buckets. Client applications can then use S3-compatible SDKs to create, list, retrieve, and delete objects on the MinIO deployment. Clients therefore drive the overall hierarchy of data within a given bucket or prefix, where Administrators can exercise control using policies to grant or deny access to an action or resource.

MinIO has no hard thresholds on the number of buckets, objects, or prefixes on a given deployment. The relative performance of the hardware and networking underlying the MinIO deployment may create a practical limit to the number of objects in a given prefix or bucket. Specifically, hardware using slower drives or network infrastructures tend to exhibit poor performance in buckets or prefixes with a flat hierarchy of objects. For other considerations, thresholds, or limitations to keep in mind, see Thresholds and Limits.

Consider the following points as general guidance for client applications workload patterns:

  • Deployments with modest or budget-focused hardware should architect their workloads to target 10,000 objects per prefix as a baseline. Increase this target based on benchmarking and monitoring of real world workloads up to what the hardware can meaningfully handle.

  • Deployments with high-performance or enterprise-grade hardware can typically handle prefixes with millions of objects or more.

MinIO SUBNET Enterprise accounts can utilize yearly architecture reviews as part of the deployment and maintenance strategy to ensure long-term performance and success of your MinIO-dependent projects.

For a deeper discussion on the benefits of limiting prefix contents, see the article on optimizing S3 performance.

Object Versioning

Changed in version RELEASE.2023-08-04T17-40-21Z: MinIO supports keeping up to 10,000 “versions” of an object in a single bucket. For workloads that require keeping more than 10K versions per object, please reach out to MinIO by email at

Object with Multiple Versions

The specific client behavior on write, list, get, or delete operations on a bucket depends on the versioning state of that bucket:


Versioning Enabled

Versioning Disabled | Suspended

PUT (Write)

Create a new full version of the object as the “latest” and assign a unique version ID

Create the object with overwrite on namespace match.

GET (Read)

Retrieve the latest version of the object by default

Supports retrieving retrieving any object version by version ID.

Retrieve the object

LIST (Read)

Retrieve the latest version of objects at the specified bucket or prefix

Supports retrieving all objects with their associated version ID.

Retrieve all objects at the specified bucket or prefix

DELETE (Write)

Creates a 0-byte “Delete Marker” for the object as “latest” (soft delete)

Supports deleting any object version by version ID (hard delete). You cannot undo hard-delete operations.

Deletes the object

See Bucket Versioning for more complete documentation.

Object Tagging

MinIO supports adding custom tags to an object. A tag is a key-value pair included in the metadata of an object. Tags can be used to control access with policies or locate an object with mc find --tags.

MinIO supports adding up to 10 custom tags to an object.

For more on setting tags, refer to mc tag set.

Object Retention

MinIO Object Locking (“Object Retention”) enforces Write-Once Read-Many (WORM) immutability to protect versioned objects from deletion. MinIO supports both duration based object retention and indefinite legal hold retention.

30 Day Locked Objects

Delete operations against a WORM-locked object depend on the specific operation:

  • Delete operations which do not specify a version ID result in the creation of a “Delete Marker”

  • Delete operations which specify the version ID of a locked object result in a WORM locking error

You can only enable object locking when first creating a bucket. Enabling bucket locking also enables versioning.

MinIO Object Locking provides key data retention compliance and meets SEC17a-4(f), FINRA 4511(C), and CFTC 1.31(c)-(d) requirements as per Cohasset Associates.

See MinIO Object Locking for more complete documentation.

Object Lifecycle Management

MinIO Object Lifecycle Management allows creating rules for time or date based automatic transition or expiry of objects. For object transition, MinIO automatically moves the object to a configured remote storage tier. For object expiry, MinIO automatically deletes the object.

MinIO applies lifecycle management rules on versioned and unversioned buckets using the same behavior as normal client operations. You can specify transition or lifecycle rules that handle the latest object versions, non-current object versions, or both.

MinIO lifecycle management is built for behavior and syntax compatibility with AWS S3 Lifecycle Management. MinIO uses JSON to describe lifecycle management rules. Conversion to or from XML may be required for importing rules created on S3 or similar compatible platforms.

See Object Lifecycle Management for more complete documentation.