Operations

Operate and maintain MinIO AIStor in production. These pages cover Day 2 and ongoing tasks, including monitoring, scaling, and troubleshooting.

Optimize your deployment with performance tuning guidelines for CPU, memory, storage, and network. Validate your configuration with benchmarking tools. Test the layers underneath with network performance testing. Run it before and after an upgrade to establish a baseline, so you can tell a regression from a bad fabric. On an InfiniBand or RoCE fabric, RDMA acceleration moves object data and erasure-coded shards without passing through the kernel network stack. When issues arise, use the debugging and troubleshooting guide for HTTP tracing, diagnostics, and metadata inspection.

Core concepts explain the foundational technologies that underpin MinIO AIStor. Erasure coding and healing cover how the cluster stores data and repairs itself. AI Workloads covers which of those mechanisms a training or inference pipeline actually touches.

MinIO AIStor provides comprehensive monitoring capabilities to maintain reliability and performance. Use metrics, logging, audit trails, and healthcheck endpoints to track deployment health and identify performance bottlenecks.

For what a bucket actually holds, bucket inventory reports produce scheduled listings of objects and their metadata. Reports are written as CSV, JSON, or Parquet, and can be filtered by prefix, age, size, name pattern, tags, or custom metadata.

As storage needs grow or change, you can scale your MinIO AIStor deployment by adding server pools to expand capacity or decommissioning pools to remove aging hardware.

For planned maintenance tasks, node maintenance tools allow you to gracefully take nodes offline without disrupting cluster operations. Cordon nodes before hardware maintenance, OS updates, or rolling upgrades.

MinIO AIStor includes tools to help you be ready in the event of hardware and software failures. The failure and recovery documentation covers handling drive failures, node failures, and disaster recovery scenarios.

MinIO AIStor requires valid enterprise licenses. The license management documentation covers the registration and renewal of the MinIO Software License used by MinIO AIStor.

Manage the TLS certificates that protect traffic to the deployment with network encryption. It covers provisioning through cert-manager, rotation on Linux and Kubernetes, and diagnosing trust chain errors.

Use batch key rotation to rotate encryption keys without disrupting ongoing operations.

Protect the data directories themselves with storage hardening, which uses SELinux or eBPF LSM to stop an administrator’s rm -rf from reaching drives the AIStor process manages.

Each MinIO AIStor release publishes release artifacts including SBOMs in CycloneDX and SPDX formats, Go module listings, and release notes. Use the REST API and RSS/JSON feeds to integrate release monitoring into existing automation pipelines. Production environments that require targeted patches without a full upgrade can request hotfix releases.

The software lifecycle policy defines end-of-life terms, the operating systems and architectures AIStor runs on, the S3 API surface it implements, and the client SDKs MinIO maintains.