Administration

Administer MinIO AIStor: manage tables and objects, configure identity and access management, set up replication, and control how your deployment operates.

Most AIStor settings are per bucket rather than per deployment, and each has its own page below. For the whole set in one place, see what you configure on a bucket.

AIStor Tables provides native support for Apache Iceberg tables directly in object storage. Create, manage, and query Iceberg tables without external catalog services or metadata databases. One deployment serves both structured tables and unstructured objects, with tables managed through the Iceberg REST Catalog API and objects through the S3 API.

Give other organizations read-only access to your tables with Table Sharing, a native implementation of the Delta Sharing protocol. Clients such as Spark, Trino, and Power BI query shared Delta and Iceberg tables with the libraries they already use.

MinIO AIStor requires authentication for every operation and uses policy-based access control to define authorized actions. Use MinIO AIStor’s Identity and Access Management (IAM) controls to manage who can access your deployment and what they can do, whether you use the built-in identity manager or integrate with external providers like OpenID Connect (OIDC) or Active Directory/LDAP.

Allow browser applications to call your deployment with CORS configuration, set globally or per bucket.

Objects and Versioning covers how MinIO AIStor organizes and manages data. Learn about managing objects, object naming conventions, and planning object hierarchies for optimal performance. Enable versioning to preserve the complete history of object changes, supporting both recovery from accidental deletions and compliance requirements.

Control object removal with locking and immutability options that enforce Write-Once Read-Many (WORM) protection on versioned objects. A GOVERNANCE or COMPLIANCE retention mode protects an object for a fixed period, and a legal hold protects it indefinitely.

Automate time-based operations on objects with Object Lifecycle Management. For example, you can expire objects after specified periods or number of versions to manage storage consumption or meet data retention policies.

Use Bucket Notifications to send event notifications to external services when operations occur on buckets or objects. MinIO AIStor supports configuring notifications for a variety of targets including Kafka, RabbitMQ, Elasticsearch, webhooks, and common database solutions.

Synchronize clusters across locations with multiple replication strategies including at the site level, at the bucket level, or using custom batch rules. Choose site replication for complete cluster synchronization, bucket replication for select buckets, or batch replication for scheduled transfers and granular control.

When several workloads share one cluster, use bucket-level Quality of Service to keep any one of them from overloading it. QoS rate-limits requests and caps concurrent operations per bucket, and its network layer shapes bandwidth before a request reaches the storage process.

Pair it with bucket quotas to cap how much data a bucket holds, and with multi-tenancy guidance to isolate teams and customers within a single deployment.

Run server-side jobs with the Batch Framework, which replicates objects, rotates encryption keys, or expires objects from a YAML job definition without moving the data through a client.

The Console UI provides a graphical user interface for managing your deployment. Access identity and access management, monitor metrics and logs, configure server settings, and browse objects through a web-based interface.

Move a deployment to a newer release with the upgrade procedures for Linux and Kubernetes. On Linux, every node restarts at the same time unless you ask for a rolling restart. That restart typically completes in under 3 seconds and S3 clients reconnect on their own, so applications generally do not notice it, and it is the mode MinIO recommends. A rolling restart keeps the cluster serving throughout but takes longer, requires a load balancer, and still reconnects the clients attached to each restarting node. See Choose a restart mode.

Migrating an existing open-source MinIO deployment to AIStor is a separate procedure with its own downtime considerations.