KubeBlocks for MinIO Overview

MinIO is a high-performance, Kubernetes-native object storage system compatible with the Amazon S3 API. KubeBlocks manages MinIO as a stateful distributed cluster with automatic credential management, health checking, and lifecycle operations.

Architecture

MinIO runs in distributed erasure coding mode where all replicas form a single pool:

Each replica is a MinIO server node
Data is striped across all nodes using erasure coding for redundancy
All nodes must be running for writes; reads tolerate up to N/2 failures
The S3 API (port 9000) and Web Console (port 9001) are exposed as services

NOTE

MinIO requires a minimum of 2 replicas. For production, use at least 4 replicas distributed across multiple Kubernetes nodes.

Supported Versions

Version	Notes
2025.10.15	Default
2024.6.29

NOTE

Starting from the 2025 release, MinIO uses source-only distribution. KubeBlocks uses the coollabsio/minio community-maintained image which provides pre-built binaries.

Features

S3-compatible API — Drop-in replacement for Amazon S3
Web Console — Browser-based management UI on port 9001
Auto credential management — Root credentials generated and stored as a Kubernetes Secret
Horizontal scale-out — Add nodes to grow storage and throughput capacity
Vertical scaling — Adjust CPU and memory
TLS encryption — Optional TLS for both API and console
Prometheus metrics — Built-in metrics at /minio/v2/metrics/cluster
Initial bucket creation — Automatically create buckets on startup via MINIO_BUCKETS env var

Limitations

Scale-in not supported — Once nodes are added, they cannot be removed. KubeBlocks blocks scale-in OpsRequests to prevent data loss.
Minimum 2 replicas per pool — Each scale-out adds a new pool; each pool requires at least 2 nodes
No parameter reconfiguration — Server configuration changes require cluster rebuild
No switchover — Distributed object store; all nodes are peers

Architecture

MinIO runs in distributed erasure coding mode where all replicas form a single pool:

Each replica is a MinIO server node

Data is striped across all nodes using erasure coding for redundancy

All nodes must be running for writes; reads tolerate up to N/2 failures

The S3 API (port 9000) and Web Console (port 9001) are exposed as services

NOTE

MinIO requires a minimum of 2 replicas. For production, use at least 4 replicas distributed across multiple Kubernetes nodes.

Version

Notes

2025.10.15

Default

2024.6.29

Features

S3-compatible API — Drop-in replacement for Amazon S3

Web Console — Browser-based management UI on port 9001

Auto credential management — Root credentials generated and stored as a Kubernetes Secret

Horizontal scale-out — Add nodes to grow storage and throughput capacity

Vertical scaling — Adjust CPU and memory

TLS encryption — Optional TLS for both API and console

Prometheus metrics — Built-in metrics at /minio/v2/metrics/cluster

Initial bucket creation — Automatically create buckets on startup via MINIO_BUCKETS env var

Limitations

Scale-in not supported — Once nodes are added, they cannot be removed. KubeBlocks blocks scale-in OpsRequests to prevent data loss.

Minimum 2 replicas per pool — Each scale-out adds a new pool; each pool requires at least 2 nodes

No parameter reconfiguration — Server configuration changes require cluster rebuild

No switchover — Distributed object store; all nodes are peers