System

Start Upgrade

Changed on

Pull the published image and hand the container recreate to a helper container.

The preflight is re-run here, so a server that became unupgradable since the caller last looked -- an executor started, another upgrade began, the daemon went away -- is refused rather than upgraded on a stale reading.

Args: request: acknowledge_executor_loss, the caller's consent to every running executor being closed as SYSTEM_CLEANUP. Required when any is running.

Returns: 202 with the run id and the initial phase. The API is replaced part-way through, so the result is read back from /system/upgrade/status after it restarts.

Raises: HTTPException: 409 when the upgrade is refused. Nothing was pulled, and nothing on the server was changed.

post/system/upgrade

Request

  • Base URL
    http://localhost:8000
  • URL
    http://localhost:8000/system/upgrade
  • Auth
    HTTP basic

Request body

acknowledge_executor_lossboolean

Confirms that every running executor will be closed as SYSTEM_CLEANUP and not restored. Required when the server has any running executor.

Response

Successful Response

unknown required

Changes