Reference: Deployment Environments

Intended audience

sysadm staff members

We have 3 deployment environments:

• staging-next-version

• staging

• production

Staging Next-version Environment

This is a sandboxed environment dedicated to run the next versions of swh (a subset of our swh modules newly released). It’s currently running in the same cluster as the staging environment (same cluster but in a dedicated namespace swh-cassandra-next-version). The backends of this environment is also kubernetes managed.

Staging Environment

Staging nodes are currently running in the hypervisor pompidou.

It is now a mix of static (renewed) bare-metal machines (backends) & vms. They are running the debian operating system. Most of which run a rancher agent and form a kubernetes cluster (archive-staging-rke2). This allows to deploy the swh stack in kubernetes.

The environment components are listed in the inventory

Provisioning source

The source for the provisioning of those nodes is declared in the swh-sysadmin-provisioning repository. Its source code is a mix of terraform with terraform-proxmox plugin DSL.

Configuration source

The source for the configuration of those nodes is our puppet manifest swh-site repository on the staging branch (for the production nodes, it’s the production branch).

Access

Those machines are ssh accessible like the production ones as long as you have vpn access to the infrastructure.

Production Environment

Production nodes are a mix of static bare-metal machines (backends) & vms. They are running the debian operating system (managed through puppet). Most of which run a rancher agent and form a kubernetes cluster (archive-production-rke2). This allows to deploy the swh stack in kubernetes.

The environment components are listed in the inventory

Todo

This section is a work in progress.