Upgrading

This chapter outlines the general procedure for upgrading Stretchy. If there are any additional version-specific steps to be performed, then those are highlighted in the change log.

Orchestrator

Stretchy Orchestrator must be shut down to perform an upgrade. However, Node Agents can continue running jobs while Stretchy Orchestrator is down.

  1. Create a backup.

  2. Stop Stretchy orchestrator using your operating system’s service manager (launchd, systemd, …​).

  3. Replace the executable binary or container image.

  4. Adjust the configuration, if necessary.

  5. Start the Orchestrator.

Node Agents should automatically reconnect after some time.

If necessary, Stretchy Orchestrator will migrate the database automatically.

Node Agent

Node Agents can be upgraded while the Orchestrator is still operating.

  1. Disable the node in the Orchestrator (go to Nodes → Edit and uncheck Enabled).

  2. Wait for any jobs that are still running on the node to complete.

  3. Stop the agent using your operating system’s service manager (launchd, systemd, …​).

  4. Replace the binary stretchy-agent.

  5. Adjust the configuration, if necessary.

  6. Start the agent and wait for it to successfully connect to the Orchestrator.

  7. Enable the node again in the Orchestrator (go to Nodes → Edit and check Enabled).