stop_engine.py

The stop_engine.py script stops all servers in the engine, starting with the processing servers and then the publishing servers. Each server is allowed to finish processing in-flight transactions before it is stopped.

Note: Before performing the stop operation, the stop_engine.py script prompts the user whether to continue because the result can cause a loss of service when the target engine is running the ACTIVE processing cluster. Enter y to continue. If the engine is the last engine in the chain and the MATRIXX deployment is designated a production environment, the script confirms a second time that the engine must be stopped. Both of these prompts are skipped if the script is run with the --no_prompt option.

For more information about configuring production environment restrictions and behavior, see the discussion about site configuration.

Syntax

stop_engine.py [-h] [--kill] [--by_blade] [-n] [--no_prompt] [-d] [--debug_config_file_parsing] [--debug_function_calls] [--debug_output_formatting][--debug_threading_code] -e engineId [-p {0|1}] [--plain] [--skip_local] [--ssh_debug] [-u username]

Options

The stop_engine.py script has the following options in addition to the general engine command line options. The script uses a --nowait flag rather than a --wait flag, as do the start and restart commands.

--kill: kill flag. When specified, forces the --by_blade option to be set and the processes are stopped by using the Linux kill process. The default is to perform an orderly shutdown by using the Cluster Manager.
--by_blade: Init.d-based shutdown flag. When specified, the engine is stopped by using OS-level /etc/init.d scripts to stop the servers individually. The default is to stop the engine by using a cluster management command to stop its clusters. In this case, the Cluster Manager initiates an orderly shutdown of all servers at one time.
-n, --nowait: When specified, initiates the shutdown of a cluster and returns immediately without waiting for the engine clusters to fully exit. The default is to wait for the cluster to exit. This option is ignored when stopping the cluster with the --by_blade option.
--no_prompt: Run the script without prompting the user to confirm whether to continue with the operation. Prompting the user prevents running the script on an engine that is running the HA ACTIVE processing cluster. Using the --no_prompt option might cause a loss of service. See the preceding note.