Rapyuta

This page explains how to use Rapyuta: The RoboEarth Cloud Engine.

For further understanding on exactly what each of these processes do and how they work please visit Developer Resources

This is the master process or the core process of the RoboEarth Cloud Engine and MUST be started first. Only 'One' Instance of this process must exist.

rce-master

Any number of these may exist, they handle robot connections to the cloud engine. In developer mode :

rce-robot master_ip

Standard deployment use case (dev_mode = False)

rce-robot master_ip adminInfra_password

Any number of these may exist, these are the client processes responsible for starting up the actual compute nodes, start One for each machine providing the possibility to launch compute environments. At the moment this requires super user privileges. In developer mode :

sudo rce-container master_ip [optional: max containers supported in the machine eg : 10]

Standard deployment use case (dev_mode = False)

sudo rce-container master_ip admin_password adminInfra_password [optional: max containers supported in the machine eg : 10]

This is a ROS client to interact with the cloud engine. Needs roscore running to work. Look at {path to rce repo}/test/debug.cfg for an example of a config file. Configuration details here.

roscore

(Run on new terminal)

rce-ros {path to config file}

This can be used to connect and administer the Users and Robots and infrastructure available to a particular user or admin.

rce-console master_ip [Optional : Port to connect, defaults to: 8081]

This will launch the prompt for username and password which the user must provide. If you are using a test installation, you must be having two usernames: admin and testUser with password same as the username and with different sets of console privileges.

• For more information on how to use the client, you could type in help on the console

help Or Go here.

Sample file available at Sample Settings

• The config file is located at ~/.rce/config.ini. Can be recreated using

  ./provision config

• The credentials are auto created and used in developer mode and located at ~/.rce/creds. Can be recreated using

  ./provision cred

  Note: Depends on settings in the config file.

The following are the defaults.

 username    : password
 admin       : admin
 adminInfra  : adminInfra
 testUser    : testUser

The default config file mounts all the contents of ~/packages into the container filesystem ROS_PACKAGE_PATH (/opt/rce/packages)

1. Place your packages in ~/packages
2. Start the container environment system

   sudo rce-make

3. Rosmake any packages you require

   rosmake <example_package>

4. Install any depends inside this environment using standard apt-get procedures as it's usually done on Ubuntu
5. Exit the environment

   exit

Use as usual

When you start a container via a client or console, it is created on a machine where a container process is running and is named as C100, C101, and so on (First created container is called C100, second created container is called C101, etc.). Considering normal installation, the IP address of the created container will be 10.0.3.100, 10.0.3.101 and so on, only internally for a machine. Externally, the IP is <IP address of machine>:8000, <IP address of machine>:8001 and so on. That's why to debug processes running on the container on the fly, you can ssh into the machine:

ssh root@10.0.3.XXX

Default password is admin, unless you changed it in the configuration.

If there is a problem with the environment process, you can check its status by running the ps command:

ps aux | grep rce-environment

If the environment process is running, to further debug you can check logs of the environment process in this file:

cat /opt/rce/data/env.log

There are other log files in the directory /opt/rce/data that are named after the running nodes and can also be helpful. One other log file rosproxy.log is related to the console client and can be ignored unless you are using the console client.