Documentation
AIStore (AIS) is a lightweight distributed storage stack for AI workloads. It can run on a single Linux machine or across production clusters, with native support for in-cluster storage, remote cloud buckets, batch jobs, SDKs, and observability.
Use this page as a quick orientation and navigation map:
- New to AIS: start with About AIStore, then follow Get Started.
- Managing data and jobs: use Use AIStore.
- Running clusters: use Operations, Observability, and Performance.
- Looking up APIs, SDKs, compatibility, security, protocols, or limits: use Reference.
- Debugging a problem: use Troubleshooting.
About AIStore
- Main README
- In-depth overview
- Terminology and core abstractions
- Networking model
- Traffic patterns
- Backend providers
- Storage services
- Buckets: design, operations, namespaces, and system buckets
- On-disk layout
- Highly available control plane
Get Started
- Getting Started
- Docker
- AIS in containerized environments
- Configuration
- Environment variables
- HTTPS and certificates
- Switching a cluster to HTTPS
Use AIStore
- CLI overview
ais help- Bucket operations
- Show cluster, bucket, and object details
- Cluster and remote-cluster management
- Storage and mountpath management
- Downloads
- Evict buckets or cached data
- Jobs
- ETL overview
- ETL CLI docs
- Archives: read, write, and list
- Downloader
- Blob Downloader
- Batch object retrieval (get-batch)
- Batch operations
- Virtual directories
- Machine learning workloads
- CLI authentication and access control
- Configuration via CLI
- GCP credentials via CLI
- TLS certificate management via CLI
Operations
- Production deployment
- AIStore on Kubernetes
- Kubernetes Operator
- Ansible playbooks
- Helm charts
- Deployment monitoring
- Node lifecycle: maintenance, shutdown, decommission
- Global rebalance
- Resilver
- Information Center (IC)
- Out-of-band updates
- Native Bucket Inventory (NBI)
- System files
Observability
- Observability overview
- Monitoring with CLI
- Logs
- Prometheus integration
- Metrics reference
- Grafana dashboards
- Kubernetes monitoring
- Distributed tracing
- Monitoring get-batch
Performance
- AIS load generator (
aisloader) - Benchmarking AIStore
- Performance tuning and testing
- Performance monitoring via CLI
- Rate limiting
- Checksumming
Reference
APIs and SDKs
- CLI reference guide
- Go API
- Python SDK
- PyPI package
- Python SDK reference guide
- PyTorch integration
- TensorFlow integration
- HTTP API reference
- curl examples
- Easy URL
Compatibility
Security and protocols
- Feature flags
- AuthN service and access control
- Authentication validation
- MessagePack protocol
- Idle connections