BlueBanquise Documentation

Welcome on BlueBanquise main documentation. You should find here all needed information on how to deploy and use the BlueBanquise stack.

Since one of the main BlueBanquise goals is also to be a training for new Linux system administrators, part of the documentation is generic knowledge and act as a tutorial for basic cluster deployment and basic Ansible usage. Advanced users can skip these parts.

These pages contain information on:

• How to manually deploy a test HPC cluster (generic knowledge, not related to the stack)

• How to use Ansible (generic knowledge, not related to the stack)

• How to bootstrap, configure, deploy and maintain BlueBanquise on different scenario clusters

• Detailed documentation on each roles provided with the stack

Please report us any issues in this documentation.

Please proceed to Introduction to start.

---

Contents:

• 1. Introduction
• 2. Vocabulary
  • 2.1. Ansible vocabulary
    • 2.1.1. Host
    • 2.1.2. Group
    • 2.1.3. Variables
    • 2.1.4. Inventory, roles, and playbooks
    • 2.1.5. Variables precedence
    • 2.1.6. Replace
    • 2.1.7. Jinja2
  • 2.2. Stack vocabulary
    • 2.2.1. Icebergs
    • 2.2.2. Equipment profiles
• 3. [Training] - Cluster SysAdmin
  • 3.1. Hardware requirements
  • 3.2. Useful commands
  • 3.3. Vocabulary
    • 3.3.1. Basic concepts
    • 3.3.2. Basic words
    • 3.3.3. Understanding services
    • 3.3.4. Computational resources management
  • 3.4. Cluster description
    • 3.4.1. Architecture
    • 3.4.2. Network
    • 3.4.3. Final notes before we start
  • 3.5. Management node installation
    • 3.5.1. Setup basic repositories
    • 3.5.2. DHCP server
    • 3.5.3. DNS server
    • 3.5.4. Hosts file
    • 3.5.5. Time server
    • 3.5.6. PXE stack
  • 3.6. Other nodes installation
    • 3.6.1. Boot over PXE
    • 3.6.2. Configure client side
  • 3.7. Storage
    • 3.7.1. NFS server
    • 3.7.2. NFS clients
  • 3.8. Slurm
    • 3.8.1. Controller
    • 3.8.2. Computes nodes
    • 3.8.3. Submitter
    • 3.8.4. Submitting jobs
  • 3.9. Users
  • 3.10. Infiniband
  • 3.11. GPU (Nvidia)
    • 3.11.1. Ensure kernel do not crash
    • 3.11.2. Disable nouveau driver
    • 3.11.3. Install Nvidia driver
  • 3.12. Conclusion
• 4. [Training] - Ansible
  • 4.1. Install Ansible
  • 4.2. Minimal inventory
  • 4.3. Ansible commands
    • 4.3.1. ansible
    • 4.3.2. ansible-inventory
    • 4.3.3. ansible-playbook
    • 4.3.4. Debug
  • 4.4. Variables and groups
    • 4.4.1. Adding variables
    • 4.4.2. Configuring groups
    • 4.4.3. Variables precedence
  • 4.5. Roles and playbooks
    • 4.5.1. Role
    • 4.5.2. Playbook
    • 4.5.3. Jinja2
    • 4.5.4. Task again
• 5. [Core] - Bootstrap base system
  • 5.1. Bootstrap RHEL like system
    • 5.1.1. OS installation
    • 5.1.2. Create bluebanquise user
    • 5.1.3. Bootstrap stack - online
    • 5.1.4. Bootstrap stack - offline
  • 5.2. Bootstrap Ubuntu like system
    • 5.2.1. OS installation
    • 5.2.2. Create bluebanquise user
    • 5.2.3. Bootstrap stack
  • 5.3. Bootstrap SUSE like system
    • 5.3.1. SLES 15
• 6. [Core] - Configure BlueBanquise
  • 6.1. Understand Environment
  • 6.2. Configure inventory
    • 6.2.1. Check example inventory
    • 6.2.2. Review groups
    • 6.2.3. Review nodes
    • 6.2.4. Review logical networks
    • 6.2.5. Review general configuration
    • 6.2.6. Review equipment default parameters
• 7. [Core] - Deploy Cluster
  • 7.1. Management deployment
    • 7.1.1. Get managements playbook
    • 7.1.2. Ansible-playbook
    • 7.1.3. Apply management1 configuration
  • 7.2. Deploy OS on other nodes: PXE
    • 7.2.1. PXE process overview
    • 7.2.2. bootset
    • 7.2.3. SSH public key
    • 7.2.4. OS deployment
  • 7.3. Apply other nodes configuration
• 8. [Core] - Manage multiple icebergs
  • 8.1. Iceberg elements
  • 8.2. Configuring top iceberg
    • 8.2.1. Understanding mechanism
    • 8.2.2. Enabling iceberg mechanism
    • 8.2.3. Create iceberg 1
    • 8.2.4. Create iceberg 2
    • 8.2.5. Deploy sub management configuration
    • 8.2.6. Interconnect and job scheduler (HPC only)
• 9. [Core] - Diskless
  • 9.1. Introduction
    • 9.1.1. Diskless
  • 9.2. Diskless Technical Documentation
    • 9.2.1. Software Architecture
    • 9.2.2. The image system
    • 9.2.3. Module creation
    • 9.2.4. Cleaning process
• 10. [Community] - High Availability
  • 10.1. Create HA Cluster
    • 10.1.1. ha_cluster group
    • 10.1.2. Reference node
    • 10.1.3. Cluster password
    • 10.1.4. Deploy cluster
  • 10.2. Populate cluster with resources
  • 10.3. Configure cluster
  • 10.4. Expanding cluster
• 11. [Community] - Monitoring
  • 11.1. Monitoring Introduction
  • 11.2. Prometheus
    • 11.2.1. Prerequisites
    • 11.2.2. Installation
    • 11.2.3. Prometheus.yml
    • 11.2.4. Variables
    • 11.2.5. Queries
    • 11.2.6. Alerts
  • 11.3. Grafana
    • 11.3.1. Installation
    • 11.3.2. Useful Files
    • 11.3.3. Dashboard
    • 11.3.4. Types of Visualization
    • 11.3.5. Extra
  • 11.4. Exporters
    • 11.4.1. Node_exporter
    • 11.4.2. Ipmi_exporter
    • 11.4.3. Snmp_exporter
    • 11.4.4. Ha_cluster_exporter
    • 11.4.5. Slurm_exporter
• 12. [External] - Deploy Kubernetes
  • 12.1. Deploy BlueBanquise CORE cluster
    • 12.1.1. Install management node
    • 12.1.2. Grab BlueBanquise
    • 12.1.3. Configure inventory
    • 12.1.4. Deploy management
    • 12.1.5. Deploy infrastructure
    • 12.1.6. Haproxy and keepalived
  • 12.2. Deploy Kubernetes cluster
    • 12.2.1. Grab Kubespray
    • 12.2.2. Configure kubespray
    • 12.2.3. Deploy K8S
    • 12.2.4. Check cluster works
  • 12.3. Install Octant
  • 12.4. Configure Kubernetes cluster
    • 12.4.1. Configure nginx-ingress together with MetalLB
    • 12.4.2. Create test resources
    • 12.4.3. Link MetalLB to world
• 13. FAQ
  • 13.1. Is it possible to put BlueBanquise into a container ?
    • 13.1.1. Podman
• 14. Stories
  • 14.1. Management and main configuration
    • 14.1.1. Versioning the inventory
    • 14.1.2. Adding a new repository
    • 14.1.3. Update management nodes configuration
  • 14.2. Nodes
    • 14.2.1. Adding a new node
    • 14.2.2. Adding a new range of nodes
    • 14.2.3. Adding a new master group
    • 14.2.4. Adding a new equipment_profile group
    • 14.2.5. Adding a custom group
    • 14.2.6. Replacing/Updating a node
    • 14.2.7. Deploying nodes
    • 14.2.8. Apply or update nodes configuration
    • 14.2.9. Changing equipment_profile group of some nodes
    • 14.2.10. Manage multiple distribution versions
  • 14.3. Roles and playbooks
    • 14.3.1. Create a custom role
  • 14.4. Security
    • 14.4.1. Update root password
    • 14.4.2. Use vault to enhance inventory security
    • 14.4.3. Externalize Ansible
  • 14.5. Storage
    • 14.5.1. LVM mirroring
• 15. Roles list
  • 15.1. CORE
    • 15.1.1. Access control
    • 15.1.2. Audit system
    • 15.1.3. ClusterShell
    • 15.1.4. Conman
    • 15.1.5. DHCP server
    • 15.1.6. Diskless
    • 15.1.7. DNS client
    • 15.1.8. DNS server
    • 15.1.9. Filesystem
    • 15.1.10. Firewall
    • 15.1.11. Hosts file
    • 15.1.12. kernel_config
    • 15.1.13. Log client
    • 15.1.14. Log server
    • 15.1.15. LVM
    • 15.1.16. Modprobe
    • 15.1.17. Filesystem
    • 15.1.18. NFS client
    • 15.1.19. NFS server
    • 15.1.20. NIC
    • 15.1.21. Pam_limits
    • 15.1.22. Parted
    • 15.1.23. Powerman
    • 15.1.24. PXE Stack
    • 15.1.25. Repositories client
    • 15.1.26. Repositories server
    • 15.1.27. Root password
    • 15.1.28. Set Hostname
    • 15.1.29. ssh client
    • 15.1.30. SSH slave
    • 15.1.31. Sudoers
    • 15.1.32. Time server
  • 15.2. ADVANCED CORE
    • 15.2.1. Advanced DHCP server
    • 15.2.2. Advanced DNS server
• 16. References
  • 16.1. External Graphics

Indices and tables

• Index

• Module Index

• Search Page

---

Documentation authors:

• LEVEUGLE Benoît

• KEATS Johnny

• TRAVOUILLON Bruno

• GELLNER Tim

• MUNDAY Neil

• RIBEIRO Adrien

• MERZOUKI Hamid

• PEROTIN Matthieu

• PIETERS David