Skip to content

Latest commit

 

History

History
114 lines (84 loc) · 6.78 KB

README.md

File metadata and controls

114 lines (84 loc) · 6.78 KB
Raw

Getting Started

This document serves as a starting guide for users to utilize the netapp_eseries.beegfs collection. It provides resources to assist users in configuring the neccesary Ansible files for a comprehensive deployment of BeeGFS within a High Availability (HA) Cluster environment.

Table of Contents

Requirements

Before following this guide, please adhere to the requirements outlined in this collection's README.

Getting Started Guide

Each role in the collection operates independently to deploy a specific version of BeeGFS, configured for a particular mode of operation. For more detailed information about a role's use case and variables, please refer to its corresponding readme file.

  • BeeGFS 7.4 with High-Availability (HA): Provides end-to-end deployment of the NetApp E-Series BeeGFS HA solution. It includes provisioning and mapping E-Series storage, creating a Linux HA cluster using Pacemaker and Corosync, deploying BeeGFS into the cluster, and configuring clients.

  • BeeGFS Client: Installs the BeeGFS Client and, if requested, mounts one or more BeeGFS file systems, or the same BeeGFS file system multiple times. The role can also unmount one or more BeeGFS file systems if requested.

To utilize these roles and begin deploying BeeGFS to storage clusters or client nodes, you will need to configure the inventory, host/group variables, passwords, and playbook files. Follow the instructions in the next section to build these files according to your BeeGFS cluster environment. The passwords.yml file requires credentials for various components in the HA structure. For a secure environment, the supplied passwords should not be stored in plain text. Use Ansible Vault to encrypt passwords.

Building Your BeeGFS HA Building Block Inventory and Playbook Files

NetApp Verified Architecture (NVA)

Standard BeeGFS configurations that are optimized to meet the performance requirements of demanding workloads which defined in NVA generations. Use one of the gen folders in beegfs_on_netapp folder to deploy a standard BeeGFS configuration for your system.

Custom Architecture

Custom configurations are defined by users. There are two methods to build your inventory and playbook files:

  • The first, and recommended method, is to follow the instructions provided in the beegfs_ha_building_block README. The instructions walk the user through creating an inventory structure, populating the inventory variables, and deploying BeeGFS HA services.

  • The second method involves manually creating the inventory, playbook, group variables, and host variables using the examples found in the beegfs_on_custom_configurations/examples/ folder. Once all the files are populated according to your configuration, BeeGFS HA services can be deployed by running the following playbook:

    ansible-playbook -i <inventory>.yml <playbook>.yml

Building Your BeeGFS Client Inventory and Playbook Files

NetApp Verified Architecture (NVA)

Use one of the gen folders in beegfs_on_netapp folder to deploy a standard BeeGFS configuration for your system.

Custom Architecture

Follow the instructions provided in the beegfs_client README. The instructions walk the user through creating an inventory, host variables, and a playbook to install the BeeGFS client service to a node and, if desired, mount one or more beeGFS file systems (or the same file system multiple times).

Example Playbook, Inventory, Group/Host Variables

Example playbook, inventory, and group/host variable files can be referenced in the beegfs_on_netapp OR beegfs_on_custom_configurations/examples folder. There you can find examples of how to layout the playbook and inventory files. In depth configuration instructions can be found on the NetApp Docs portal: BeeGFS on NetApp with E-Series Storage.

The variables used in the inventory files are not exhaustive, see additional variables from under role variables or other NetApp E-Series Ansible Collections (santricity, host).

General Notes

  • All BeeGFS cluster nodes must be available for the deployment process to function properly.
  • Fencing agents should be used to ensure failed nodes are definitely down.
  • BeeGFS is added to the PRUNEFS list in /etc/updatedb.conf to prevent daily indexing scans on clients which causes performance degradations.
  • Please refer to the documentation for your Linux distribution/version for guidance on the maximum cluster size. For example, the limitations for RedHat can be found here.

Maintainer Information

  • Christian Whiteside (@mcwhiteside)
  • Vu Tran (@VuTran007)