It's error prone to start/stop the easegress executable binary directly in a multi-node easegress group. This is because the embedded etcd cluster may be corrupted if one peer node starts more than one instances.

This sbin script suite specifies and regulates the basic directory structure, config file path, binary path, and provide start.sh and stop.sh to assist the admin does not start/stop the easegress process in any wrong way.

1. Let's define the APPDIR is the absolute path to the directory containing start.sh and stop.sh.
2. Copy the executable binary easegress-server and egctl into $APPDIR/bin.
3. Edit $APPDIR/conf/config.yaml, to specify the starting parameters. Do not change the path and filename. In eg.yaml, the ./ directory is relative to APPDIR, instead of the location of eg.yaml.
4. Run start.sh to launch the new easegress process. start.sh will ignore if there has been an instance running.
5. Run stop.sh to terminate the process. stop.sh will warn and quit if there's not enough etcd members alive. run stop.sh -f if you really want to stop the node.
6. Run status.sh to check the overview status of the easegress group. Run status.sh -l to check the detailed status of the easegress group.
7. Run egctl.sh with parameters, this will delegate to APPDIR/bin/egctl and send request to the API server defined by api-addr in $APPDIR/conf/eg.yaml