MinIO

This skill covers MinIO S3-compatible object storage, including:

• Erasure coding (Reed-Solomon), drive quorums, and inline bitrot protection
• Server pool architecture, erasure set sizing, and horizontal scaling
• mc client for S3 operations and mc admin for cluster management
• KES key encryption service and SSE-S3/SSE-KMS/SSE-C modes
• Site replication (active-active multi-site), bucket replication, and batch operations
• ILM lifecycle policies, tiering, and object expiration
• IAM (built-in users, LDAP, OIDC, STS) and AWS-compatible policy engine
• MinIO Operator for Kubernetes with Tenant CRD management
• Performance tuning for NVMe, AVX-512, and high-throughput workloads
• Licensing landscape (AGPLv3 archived community vs AIStor commercial)

For cross-platform storage comparisons, see the overview skill.

How to Approach Tasks

1. Classify the request:

   • Troubleshooting -- Load references/diagnostics.md for drive failures, healing, performance benchmarks, replication lag, and common issues
   • Architecture / design -- Load references/architecture.md for erasure coding, storage hierarchy, request pipeline, KES, site replication, and Operator
   • Best practices -- Load references/best-practices.md for hardware sizing, EC selection, bucket design, security hardening, K8s deployment, and backup strategy

2. Assess edition -- Determine if the user runs archived community edition or AIStor. Key differences: community (AGPLv3, archived Feb 2026, no patches) vs AIStor Free/Enterprise (active, proprietary/commercial).

3. Load context -- Read the relevant reference file.

4. Analyze -- Consider erasure set size, drive count, quorum rules, network topology.

5. Recommend -- Provide actionable guidance with mc commands and configuration examples.

Core Architecture

Storage Hierarchy

Server Pools
  └── Erasure Sets (4-16 drives, up to 32 in AIStor 2026+)
        └── Physical Drives (NVMe recommended)

• Each pool manages its own erasure sets independently
• Pool expansion is the only supported horizontal scale-out path
• Object placement: hash(object_key) -> erasure set index

Erasure Coding

For each write: object split into K data shards + M parity shards, written in parallel. Default: K = M = N/2 where N is drives per erasure set.

Operation	Minimum Drives Required
Write	(N/2) + 1
Read	N/2

AVX-512 SIMD acceleration for near wire-speed erasure coding. Inline bitrot protection via HighwayHash-256 on every shard.

Request Pipeline

Client HTTP/S -> TLS -> Auth (IAM) -> S3 API Handler -> ObjectLayer
  -> ErasurePoolsObjects -> ErasureObjects -> Physical Drives

Key Capabilities

Feature	Detail
S3 API compatibility	Near-complete AWS S3 API coverage
Object versioning	UUID version IDs, delete markers, noncurrent version management
Object locking (WORM)	COMPLIANCE and GOVERNANCE modes
Server-side encryption	SSE-S3, SSE-KMS (via KES), SSE-C
Site replication	Active-active multi-site, IAM + data + metadata sync
ILM / Lifecycle	Expiration, tiering/transition to warm/cold targets
Batch operations	Server-side replicate, expire, keyrotate via YAML jobs
Bucket notifications	Kafka, AMQP, NATS, PostgreSQL, Redis, Webhook, and more
Kubernetes Operator	Tenant CRD, auto-TLS, declarative management

Licensing (April 2026)

Edition	Status	License
Community (OSS)	Archived Feb 13, 2026	AGPLv3
AIStor Free	Active	Proprietary
AIStor Enterprise	Active	Commercial (~$96K/year)

The archived community repository remains functional but receives no patches.

Reference Files

• references/architecture.md -- Erasure coding, storage hierarchy, request pipeline, KES, site replication, Operator, notifications, ILM, IAM
• references/best-practices.md -- Hardware sizing, EC selection, bucket design, security hardening, K8s deployment, performance tuning, backup strategy
• references/diagnostics.md -- mc admin commands, health checks, drive failures, healing, performance benchmarks, network diagnostics, common issues