Hello,

      I tried to install rapyuta on local linux host (14.04). Rapyuta neither was updated for 14.04 nor for ROS indigo. Please solve this issue. Thanks

• support for automatic connection

clean up and properly document the command line tools.

I got this problem...At first, i thought the system can't fing find the srv name, so i tried
$rossrv list

Test/StrintEcho and the 3 others service was in the list:

Test/ConverterTest
Test/ParameterTest
Test/StringEcho
Test/StringTest

I'm so confuse...Thanks for help SOS

The Install instructions need fixing , when followed word for word, something makes it say 'ResourceNotFound: Can not load manifest for ROS package "Test".' my guess is environment variables .Same bug report from the folks at Zaragoza

When calling rce-ros with certain configuration files, I get a number of Received error message: Fatal Error. On the master console, there is no error shown, but on the robot console I get a corresponding number of long stack traces ending with KeyError: 'odom', with various keys.

The curious thing is that with exactly the same .cfg file, I get these errors or not only depending on the order in which I define interfaces, connections and their properties.

As a simple example,

"interfaces" : [
        {
            "eTag"  : "cTag_01",
            "iTag"  : "cmdVel", 
            "iType" : "SubscriberInterface",
            "iCls"  : "geometry_msgs/Twist", 
            "addr"  : "youbot/cmd_vel"
        }, {
            "eTag"  : "youbot",
            "iTag"  : "cmdVel", 
            "iType" : "PublisherConverter",
            "iCls"  : "geometry_msgs/Twist", 
            "addr"  : "youbot/cmd_vel"
        }, {
            "eTag"  : "cTag_01",
            "iTag"  : "odom", 
            "iType" : "PublisherInterface",
            "iCls"  : "nav_msgs/Odometry", 
            "addr"  : "youbot/odom"
        }, {
            "eTag"  : "youbot",
            "iTag"  : "odom", 
            "iType" : "SubscriberConverter",
            "iCls"  : "nav_msgs/Odometry", 
            "addr"  : "youbot/odom"
        }
    ],

    "connections" : [
        {
            "tagA" : "cTag_01/cmdVel",
            "tagB" : "youbot/cmdVel"
        }, {
            "tagA" : "cTag_01/odom",
            "tagB" : "youbot/odom"
        }
    ]

doesn't give any errors, while

"interfaces" : [
       {
            "eTag"  : "cTag_01",
            "iTag"  : "odom", 
            "iType" : "PublisherInterface",
            "iCls"  : "nav_msgs/Odometry", 
            "addr"  : "youbot/odom"
        }, {
            "eTag"  : "youbot",
            "iTag"  : "odom", 
            "iType" : "SubscriberConverter",
            "iCls"  : "nav_msgs/Odometry", 
            "addr"  : "youbot/odom"
        }, {
            "eTag"  : "cTag_01",
            "iTag"  : "cmdVel", 
            "iType" : "SubscriberInterface",
            "iCls"  : "geometry_msgs/Twist", 
            "addr"  : "youbot/cmd_vel"
        }, {
            "eTag"  : "youbot",
            "iTag"  : "cmdVel", 
            "iType" : "PublisherConverter",
            "iCls"  : "geometry_msgs/Twist", 
            "addr"  : "youbot/cmd_vel"
        }
    ],

    "connections" : [
       {
            "tagA" : "cTag_01/odom",
            "tagB" : "youbot/odom"
        }, {
            "tagA" : "cTag_01/cmdVel",
            "tagB" : "youbot/cmdVel"
        }
    ]

prints a few errors (3 to 4).

When I try the config file with all the topics (joint, sensors, gripper) I get many more errors. In general anyway, everything works fine after the errors (that is, echoing the topics on the container works fine).

Benchmark data throughput between the old re_comm and the new fake_re_comm with hdfs-fuse.

is there a way to install rapyuta with ros kinetic which installed on ubuntu 16.04 LTS ?
as it not listed in the options while installing the cloud

Documentation: Mapping on the Cloud (ROSCon2013)

Create a Wiki Entry about the features and options offered by these
@gajanidsc please advice on where in the wiki you would like these ?

Read via a documention from the perspective of a new PhD in mechanical engineering and report to the Group.

Update Data flow in Developer Resources.

Push rce packages to PyPI and allow setup of packages as per choice.

Investigate (no implementation) on the methods to create a private virtual network between a specific set of containers (including the case where the set resides across multiple machines).

Task: Investigate and present the findings Wed (29. May) 17:00.

@dhananjaysathe: Please put this top priority for this week.

@ http://rapyuta.org/Maintainer_Resources

Add a script/tool to list info from config file, eg

1. Where is container filesystem (maybe so more useful info regarding the filesystem?)
2. Which package paths are mounted in the container
3. Gather a list of ROS packages available in the container (based on the package paths mounted in the container)
4. print a list of users in the cred db

Maybe add a feature to fetch all log files to simplify bug reports (and any other info we usually need)? (This implies that the logs are no longer printed to stdout, but to an actual logfile...)

anything else?

From this discussion I discovered that this issue is related to not shutting down correctly the container.

This is what I try to do (following the tutorial):

• add the path to test in the config file
• sudo rce-make + rosmake Test
• rce-master
• sudo rce-container localhost
• rce-robot localhost
• rce-ros debug.cfg

At this point I get various There is already a configuration directory for 'C100'.. After this error, trying to shut everything down and restarting again gives a OSError: [Errno 17] File exists: '/opt/rce/container/rootfs/opt/rce/packages/test' Could not setup the rce environment.

Add Virtual Networking Support. Hello graphLab!

Add a documentation/tips on debugging inside the container. Starting from finding the container ip address.

Make Vlad happy by adding all sourcing to .profile.

For the Getting Started page, an introduction.

Add a fixed width buffer on the client and the server side.

Sync the header with the binary blob? (@Sathe: Not sure if we were planing this for this week. Please update)

Python nodes are already tested by Dominique. Test with C++ nodes and see what is the issue with Vlad.

** This comes after the buffering issue **

Run benchmark from supposedly a robot process running on a machine to a ROS environment running on amazon via ROSbridge.

In case one wants to run only rce-master & rce-robot to create a network of robots (-> there is no need to create a container filesystem at all, ie run only provision config) then the execution of the rce processes fails because the container filesystem does not exist. (Checked in settings parser.) Some rethinking of the configuration is necessary, either:

1. create individual files for each process (and one common for all processes?)
2. add some kind of 'necessary capability' flag to indicated whether the container filesystem is required (similar comm ports for WebSocket is only necessary for rce-robot)
3. move the validity checks out of the settings; but there was a reason why they were added there in the first place...

• Add CUDA as special feature (?) for a machine
• Add CUDA setup in container to provision script as an (optional) feature
• Let the provision script automatically detect whether CUDA feature is feasible?
• Necessary to extend lxc config file with: > lxc.cgroup.devices.allow = c 195:* rwm

Open problem is the NVIDIA driver: I guess host and container driver version have to match?

How to setup CUDA for a container see: https://github.com/IDSCETHZurich/rapyuta-kinfu

a. Create AMI post milestone 2
b. Update http://rapyuta.org/EC2_Public_AMI
c. Release to marketplace

a complete

Problem: Launching of roslisp nodes can be slow.
Reason: The nodes are compiled in every container anew because the lisp interpreter 'sbcl' compiles the code and stores the resulting "binaries" under '~/.cache/common-lisp' which is a temporary folder.

Write Unit Tests, Fix Documentation

If you put a non-existing path in machine/packages and run sudo rce-make, you simply get 'Could not setup the rce-environment'.

1. Update/add the simple test case with string_msg
2. Add the bar code detection as a tutorial [If there is a time constrain only the barcode extraction part should suffice]

Mention about the Wiki in USAGE/README. Should we have a READ_ME_FIRST file? As a first point of entry?

Pull the new autobahn updates into rce_autobahn

Use GraphLab, EC2 to parallelize Global Optimization on SLAM. *Parallelize

It's not possible to publish same topic into container and subscribe to it simultaneously. For example I want to publish visual odometry transformations into /tf topic from the robot and map to odometry correction transformation into /tf from the container. Now it is not possible.

It is usually a good idea to have the /tf topic the same across all machines, so everybody can read and write to it.

Allow storage of data to RoboEarth through the cloud engine.

In case a ServiceProvider is called which does not have a connection (situation occurs if debug config file is run with rce-ros and the a rosservice call is made before the EnvironmentEndpoint is ready & connected to the RobotEndpoint) the request is lost and is never processed, i.e. the client waits indefinitely for a response. (Applies to all interfaces, but only in case of the ServiceProvider this is really a problem.)

Possible solutions:

• Might be fixed when buffering is added
• Only create client-side ROS interfaces when instructed by the RobotEndpoint, i.e. only if the there is a connection for the ServiceProvider (this is already implemented on the container side)

Fully testedRoboEarth storage that can be used as a standalone. A lot of folks are interested in using it. Should be robust and well documented.

Add a Terminology which defines the Rapyuta specific terminologies. File name suggestion [TERMINOLOGY]

• Should we have a documentation folder in the repository?

On how to generate and update the APIs.
Please add them to http://rapyuta.org/Maintainer_Resources