waldur / waldur-site-agent

An agent for integrating service provider offerings with Waldur

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Agent for Service Provider Integration

Agent for Mastermind integration with a provider's site. The main purpose of the agent is data syncronization between Waldur instance and an application (for example SLURM or MOAB cluster). The agent uses order-related information from Waldur to manage accounts in the site and accounting-related info from the site to update usage data in Waldur. For now, the agent supports only SLURM cluster as a site.

Architecture

This is a stateless application, which is deployed on a machine having access to SLURM cluster data. The agent consists of three sub-applications:

  • agent-order-process, which fetches ordering data from Waldur and updates a state of a backend correspondingly; (e.g. creates/updates/deletes SLURM accounts);
  • agent-report, which reports computing usage and limits info from a backend to Waldur (e.g. update of resource usages);
  • agent-membership-sync, which syncs membership info between Waldur and a backend (e.g. adds users to a SLURM allocation).

Integration with Waldur

For this, the agent uses Waldur client based on Python and REST communication with Waldur backend. Agent-order-process application pulls data of orders created in Waldur and creates/updates/removes backend resources based on this info. Agent-report fetches usage data pushes it to Waldur. Agent-membership-sync fetches associations from a backend and syncronizes it with remote ones.

Integration with the site

SLURM cluster

The agent relies on SLURM command line utilities (e.g. sacct and sacctmgr) and should run on a headnode of the SLURM cluster.

Agent configuration

The application supports the following CLI arguments:

  • -m, --mode - mode of agent; supported values: order_process, report and membership_sync; default is order_process.
  • -c, --config-file - path to the config file with provider settings.

Optional environment variables for dependencies:

  • REQUESTS_VERIFY_SSL - flag for SSL verification for Waldur client, default is true.
  • SENTRY_ENVIRONMENT - name of the Sentry environment.

The primary config for the agent is a waldur-site-agent-config.yaml. Using it, the agent can serve several offerings and setup backend-related data, for example computing component settings. File example and reference.

Deployment

A user should deploy 3 separate instances of the agent. The first one (called agent-order-process) fetches data from Waldur with further processing, the second one (called agent-report) sends usage data from the backend to Waldur and the third one syncs membership information between Waldur and the backend. All the instances must be configured with CLI variables and provider config file.

To deploy them, you need to setup and start the systemd services.

Prerequisite: offering configuration in Waldur

SLURM

The agents require existing offering data in Waldur. As a service provider owner, you should create an offering in the marketplace:

  • Go to Service Provider section of the organization and open offering creation menu
  • Input a name, choose a category, select SLURM remote allocation from the drop-down list on the bottom and click Create button

offering-uuid

  • Open the offering page and create a plan in the Accounting section: click Add plan and input the necessary details
  • Go to Integration section, click Show integration steps and ensure they are completed within your SLURM cluster.

Setup

Firstly, install the waldur-site-agent module:

pip install waldur-site-agent

Secondly, create the provider config file and adjust the content for your needs.

cp examples/waldur-site-agent-config.yaml.example /etc/waldur/waldur-site-agent-config.yaml

Please use the waldur_site_load_components command to load computing components into Waldur.

waldur_site_load_components

Thirdly, put systemd unit and provider config files to the corresponding locations.

# For agent-order-process
cp systemd-conf/agent-order-process/agent.service /etc/systemd/system/waldur-agent-order-process.service

# For agent-report
cp systemd-conf/agent-report/agent.service /etc/systemd/system/waldur-agent-report.service

# For agent-membership-sync
cp systemd-conf/agent-membership-sync/agent.service /etc/systemd/system/waldur-agent-membership-sync.service

After these preparation steps, run the following script to apply the changes.

systemctl daemon-reload
systemctl start waldur-agent-order-process.service
systemctl enable waldur-agent-order-process.service # to start after reboot
systemctl start waldur-agent-report.service
systemctl enable waldur-agent-report.service
systemctl start waldur-agent-membership-sync.service
systemctl enable waldur-agent-membership-sync.service

Older systemd versions

If you want to deploy the agents on a machine with systemd revision older than 240, you should use files with legacy configuration:

# For agent-order-process
cp systemd-conf/agent-order-process/agent-legacy.service /etc/systemd/system/waldur-agent-order-process-legacy.service
# For agent-report
cp systemd-conf/agent-report/agent-legacy.service /etc/systemd/system/waldur-agent-report-legacy.service
# For agent-membership-sync
cp systemd-conf/agent-membership-sync/agent-legacy.service /etc/systemd/system/waldur-agent-membership-sync-legacy.service

Provider config file reference

sentry_dsn: "" # Data Source Name for Sentry (more info https://docs.sentry.io/product/sentry-basics/dsn-explainer/).
offerings: # Settings for offerings
  - name: "Example Offering" # offering name
    waldur_api_url: "http://localhost:8081/api/" # URL of Waldur API (e.g. http://localhost:8081/api/).
    waldur_api_token: "" # Token to access the Waldur API.
    waldur_offering_uuid: "" # UUID of the offering in Waldur.
    backend_type: "slurm" # type of backend, for now only `slurm` is supported
    backend_settings: # backend-specific settings
      default_account: "root" # Default parent account name in SLURM cluster
        # for new ones
      customer_prefix: "hpc_" # Prefix for customer's accounts.
      project_prefix: "hpc_" # Prefix for project's accounts.
      allocation_prefix: "hpc_" # Prefix used for allocation's accounts.
      allocation_name_max_len: 34 # Maximum length of account name
        # created by the agent.
      enable_user_homedir_account_creation: true # Whether to create home directories
        # for users associated to accounts.
    backend_components: # Computing components on backend with accounting data
      cpu: # Type of the component, for example `cpu`
        limit: 10 # Amount of measured units for Waldur (SLURM measured unit is CPU-minutes)
        measured_unit: "k-Hours" # Waldur measured unit for accounting.
          # For example `k-Hours` for CPU
        unit_factor: 60000 # Factor for conversion from measured unit
          # to backend ones.
          # For example 60000 (60 * 1000) for CPU in SLURM,
          # which uses cpu-minutes for accounting
        accounting_type: "usage" # Can be either `usage` or `limit`
        label: "CPU" # A label for the component in Waldur
      mem:
        limit: 10 # Amount of measured units for Waldur (SLURM measured unit is MB-minutes)
        measured_unit: 'gb-Hours' # Waldur measured unit for accounting
        unit_factor: 61440 # Unit factor for conversion from measured unit
          # to SLURM units (60 * 1024)
        accounting_type: usage # Can be usage or limit
        label: RAM # A label for a component in Waldur

About

An agent for integrating service provider offerings with Waldur

License:MIT License


Languages

Language:Python 100.0%