Following is a high-level view of the nature and order of tasks that need to be performed for a full-stack Swarm implementation.

Phase 1: Prepare Environment

The work to prepare the environment must be completed before adding any Swarm components:

• Rack and stack hardware designated for Swarm, replacing and upgrading as needed. (See Hardware Setup)

• Upgrade firmware to the latest versions:

  • All servers

  • All disk controllers

  • All disk drives

• Configure networking and switches (see Network Infrastructure), including the following:

  • VLAN configuration

  • IGMP snooping disabled (or IGMP querier implemented)

• Configure IPMI management.

• Provide access for the storage cluster to phone home. (See Health Data to Support)

• Verify the servers and base operating systems meet the Swarm system requirements. (See Hardware Requirements for Storage and Hardware Requirements for Elasticsearch Cluster)

• Configure IPMI (remote server management)

• Complete licenses and agreements

  • Obtain any needed storage capacity and capability licenses from DataCore. (See Licensing Swarm)

  • A user in the organization must register the Red Hat license and accept the EULA if installing or updating RHEL.

  • Accept the DataCore EULA.

Phase 2: Platform Server and Storage Cluster

Swarm Platform Server is installed and configured first so that the platform server can install Storage nodes on the designated hardware.

• Install Platform Server. (See SCS Installation)

• Configure Platform Server to integrate with the environment.

• Configure Platform Server to boot the current version of Swarm Storage.

• Boot the Storage nodes and configure the cluster-wide settings. (See Configuring Swarm Storage)

• Install the Swarm Storage UI. (See Swarm Storage UI Installation)

• Verify the storage cluster is operational: read, write, and delete test objects

• Optional: Install the open-source components to use Swarm's exports to Prometheus. (See Prometheus Node Exporter and Grafana)

Phase 3: Elasticsearch

Install and configure an Elasticsearch cluster on designated hardware, providing the Storage cluster with the search and metrics capabilities.

• Base install the chosen operating system (RHEL/CentOS 7 or RHEL/Rocky Linux 8).

• Install Elasticsearch nodes on designated hardware. (See Elasticsearch Implementation)

• Configure Elasticsearch based on DataCore recommendations. (See Configuring Elasticsearch)

• Create a Search feed to populate the Elasticsearch metadata index. (See Managing Feeds)

• Configure Elasticsearch Curator and Swarm Metrics. (See https://perifery.atlassian.net/wiki/spaces/DOCS/pages/2443810001)

Phase 4: Content Gateway

Install and configure Content Gateways, which provide the primary access to the Storage cluster.

• Base install the chosen operating system (RHEL/CentOS 7 or RHEL/Rocky Linux 8).

• Install the Content Gateway. See https://perifery.atlassian.net/wiki/spaces/public/pages/2443810075

• Install the Content UI. See https://perifery.atlassian.net/wiki/spaces/public/pages/2443810310

• Configure basic gateway setup for verification and initialization of primary domain. See https://perifery.atlassian.net/wiki/spaces/public/pages/2443810201 and https://perifery.atlassian.net/wiki/spaces/public/pages/2443810131

• Verify the Gateway is operational: read, write, and delete test objects using the Content UI, S3, and SCSP.

• Create the initial domains, with policy definitions. See https://perifery.atlassian.net/wiki/spaces/public/pages/2443816967

Phase 5: Swarm Clients

As fits with the implementation plan, extend access to Swarm storage by installing one or more Swarm client applications, such as the following:

• SwarmFS Implementation

• FileFly

Phase 6: Post-Installation

• Conduct performance measurement and tuning.

• Test and debug third-party/custom applications and integration.

• Train administrators and staff.