September 2013 - Ibr-dtn - lists.ibr.cs.tu-bs.de

OpenWrt ibr timestamp option
by Dimitrios Giotas 30 Sep '13

30 Sep '13

Hello, can someone tell me how to enable the --timestamp option in ibr-dtn in OpenWrt so I can see the unix time in the log file? Thank you. Regards, Dimitris

2 1

Getting rid of annoying Time-sync on different nodes
by Shyam B 23 Sep '13

23 Sep '13

Hello All, I am working on a project with DTN in which while exchanging the bundles either don't reach the destination or the file transfer fails completely. Then I routed the problem on different times on the system. There was some google page about BAD_Clock, but I am not able to find that page on ibrdtn wiki anymore. Could someone help me with, how to configure: 1) Considering completely BAD_Clocks on different nodes, I can still be able to transmit data. I don't mind time conscience/ security in this regard, for now. My nodes (RPis) are not connected to internet, so they don't know anything about time if they are switched off. Moreoever, I do not have a real-time clock for it. 2) When setting time_synchronize parameter in the ibrdtnd.conf file does not help. What does this really do, any documentation on this and other <time> related parameter would be helpful. Thanks a lot! -- Best Regards, Shyam

2 1

IBR-DTN doesn't find neighbors on Android Emulator
by Alejandro Willers 19 Sep '13

19 Sep '13

Hi, I am testing IBR-DTN on Android Emulator and I can't find any neighbors in spite of having an Android Phone on the proximity. Is there any limitation why IBR-DTN should not work on Emulator? Thanks & Regards

2 2

[RESOLVED] DTNTrigger not working!
by Shyam B 19 Sep '13

19 Sep '13

Hi, I downloaded the ibrdtn code and debugged dtnrecv and dtntrigger step by step to understand what's going on. Finally, I know what was the cause. I was at fault, though I won't call it my mistake ;) dtnrecv always worked when the other did not, then I found that the <name> parameter was different at both ends. dtn::data::Bundle b = client.getBundle(); gets hung because client initialization gets an incorrect group <name> for the bundle. It must be the same as that of the bundle sent. Don't know why so much security, we already have time-sync, encryption options? Curious, why it is designed this way <name>? One Request: ------------- dtnsend -h [brings about the following:] --src <name> set the source application name (e.g. filetransfer) dtnrecv -h [brings about the following:] --src <name> set the source application name (e.g. filetransfer) dtntrigger -h brings about the following:] <name> the application name May I request to add to the print_help() in .cpp items more information in the next release about having the 'same' <name> on both sender and receiver, else it will not work as desired! It may look intuitive to developers but possibly counter-intuitive to users. An example of how to use the commands (under --help) will certainly go a long way (I know some is there in wiki, but still).. To add it to the print_help() cout << "- - - - - \nExample usage at receiving end:\n"<<"\tdtntrigger filetransfer python /full-path-to-file\n"<<"Example usage for this at sending end:\n"<<"\tdtnsend dtn://target_node/filetransfer somefile.mp3"; Thanks all, but I certainly did learn more for this day :) Now, this makes me to invent some features in a new tool I would call 'dtnfire' :) Whew!! PS: Been a s/w developer for 7 years now, I must say, very well written and commented code!! -- Best Regards, Shyam

1 0

DTNTrigger not working: Help! (UPDATE - 2)
by Shyam B 18 Sep '13

18 Sep '13

Dear Johannes/*Stephan*, Thank you for the reply to my post. The code had a type '&', which I did not intend to run in background :) The ::system() call works (as I see in eclipse), the 'complete script path' did the trick, however, the DTNTRIGGER.exe does not work still work. I wonder, where I am going wrong. Updated code: --------------------- #include <csignal> #include <ctype.h> #include <stdio.h> #include <stdlib.h> #include <unistd.h> //#include "/home/s116793/Dropbox/AmazingCode/ScriptTest/dtncode/ibrdtn-0.10.1/ibrdtn/data/Bundle.h" #include <ibrdtn/data/Bundle.h> #include <iostream> using namespace std; // std::string _appname = "trigger"; std::string _script = "/home/s116793/bashscript.sh"; std::string _shell = "/bin/sh"; int main() { cout << "!!!Hello World!!!" << endl; // prints !!!Hello World!!! // call the script std::string cmd = _shell + " " + _script + " " + "NodeA" + " " + "/home/s116793/something"; cout << "The cmd is:"<<cmd<<endl; ::system(cmd.c_str());; return 0; } OUTPUT: -------------- !!!Hello World!!! The cmd is:/bin/sh /home/s116793/bashscript.sh NodeA /home/s116793/something Hello guys.. Number of arguments 1 arguments Argument list: ['/home/s116793/triggered.py'] Name of the file /home/s116793/text.txt ------------------- I did try what you suggested, giving an app-name for the dtntrigger: the command at the receiver: dtntrigger ubuntu-app /bin/sh /home/s116793/bashscript.sh No file is created! Another thing to notice: -------------------------------- When I look into the cpp code also, we have set an app_name = "trigger" which is overwritten in the switch case. ONE minor correction: // print help if not enough parameters are set 125 if (optindex < 2) { print_help<http://www.ibr.cs.tu-bs.de/projects/ibr-dtn/doxygen/0.10.0/dtnconvert_8cpp.…>(); exit(0); } This should be "optindex <= 2", forcing the user to enter three command line arguments. Entering two arguments with dtntrigger call, sets the _appname and _shell and setting _script is left out. -- Best Regards, Shyam

3 2

DTNTrigger calling script: Help! (UPDATE)
by Shyam B 18 Sep '13

18 Sep '13

Update to previous post: ------------------------ I wrote a little C++ program with a system() call, calling the script without DTNTrigger to see if ::system(cmd.c_str()) call works in the first place. (Procedure exact same as in the dtntrigger.exe). I have tested this on both ubuntu and RaspberryPI (ARM based) #include <csignal> #include <ctype.h> #include <stdio.h> #include <stdlib.h> #include <unistd.h> #include <iostream> using namespace std; std::string _script = "bashScript.sh &"; std::string _shell = "/bin/sh"; int main() { cout << "!!!Hello World!!!" << endl; // call the script std::string cmd = _shell + " " + _script + " " + "Node A" + " " + "/home/s116793/something"; cout << "The cmd is:"<<cmd<<endl; ::system(cmd.c_str()); return 0; } Script contents: ----------- #!/bin/sh echo "Hello guys.." python /home/s116793/triggered.py Output ------ !!!Hello World!!! The cmd is:/bin/sh bashScript.sh & Node A /home/s116793/something sh: 1: Node: not found /bin/sh: 0: Can't open bashScript.sh Observations: ------------ 1) The bashscript.sh is in the same path as the executable in /Debug and it has 'full' permissions (chmod 777). I tried googling it seems to be a known problem for different reasons, not sure what's the problem here. Any expert advice? 2) Something wrong with the way I call? Can someone run this on their system and post some example here? Thanks.

2 2

DTNTrigger not working: Help!
by Shyam B 18 Sep '13

18 Sep '13

Hi, I am trying to run *dtntrigger*, but it is *not working*. Its been many hours since I am trying every way to get it to work! This is what i am executing on the receiving end: (Lets call this (**) for later) dtntrigger /bin/sh /home/pi/bashscript.sh -- *Contents* of bashscript.sh: #!/bin/sh echo "Hello guys.." python /home/pi/triggered.py -- *Contents* of triggered.py: (Just creating a file for verification, that when a bundle is received this be created:) #!/usr/bin/python import sys import os def main(argv): print 'Number of arguments', len(sys.argv), 'arguments' print 'Argument list:', str(sys.argv) fo = open('/home/pi/text.txt', 'w') fo.write('This is my first file') print 'Name of the file', fo.name fo.close() if __name__ == "__main__": main(sys.argv[1:]) -- I see from dtntrigger.cpp, this simple call is done: std::string cmd = _shell + " " + _script + " " + b.getSource.getString() + " " + file.getPath ::system(cmd.c_str()) -- Observations: 1) If I execute just this command "/bin/sh /home/pi/bashscript.sh" without dtntrigger, on my command line, the script executes and creates that 'text.txt' dummy file. While, with dtntrigger it does not. 2) Permissions? Yes, I have given full permissions (chmod 777) to both scripts and as you see given the full file's absolute path as well. 3) When I execute the command (**) it runs (does not break) means __running in C++ code is true. What is wrong then? 4) The bundle is available in /bundles folder (as per my ibrdtnd.conf), bundle is always coming there. I am using raspberryPI. dtnrecv works directly, no problem. I have also used ubuntu, same results!! 5) Can I execute a python script directly instead of going through bash? 6) I doubt ::system(..) call is somehow not working. Running raspbian on PI. Any help will be much appreciated to get this up and running! -- Best Regards, Shyam

2 1

IBR-DTN release 0.10.1
by Johannes Morgenroth 16 Sep '13

16 Sep '13

Patch release 0.10.1 is available! This release do not introduce any new features. Instead it is dedicated to fix bugs which could lead to segmentation faults. At this point, I would like to say 'Thanks!' to the community on the mailing list, which was really helpful in locating the issues. For a full list of features and changes watch the official changelog: http://www.ibr.cs.tu-bs.de/trac/ibr-dtn/wiki/changelog The packages for Debian, Raspbian and OpenWrt should be available right now. Kind regards, Johannes Morgenroth

1 0

DTNTrigger and Statistical APIs
by Shyam B 11 Sep '13

11 Sep '13

Dear Sebastian, Johannes, Thank you for the quick response. For my questions earlier related the earlier post here<http://www.ibr.cs.tu-bs.de/pipermail/ibr-dtn/2013-September/000472.html> : >* 1. How will 'D' ever know 'who' sent him the bundle. Currently, it just shows some weird name of the file that was received. If I don't run DTNRecv --file 'name', it is stored with weird filename. Was it A, B or C who sent this message? Is there a way for 'D' to find this out?* @Sebastian: I will write a script to process the DTNTrigger event as you suggest. Moreover, you said if I can use the IBR-DTN API directly though I haven't found information when I searched. Can you/ someone tell me where can I get information on this API, is this an API I can use from my own python script or an external API I can register as an event in my own script? *2. Statistical logging* @Sebastian/Johannes: Thank you for the interesting script. I had a quick look. Connects to the localhost and retrieves the neighbourlist. Interesting. Let me implement this and be the first thing I do at work tomorrow morning :) @Johannes: Sebastian said, *there are other kinds of statistics that can be obtained from the API, but I am not sure they are document*. Can you please help me or point me to that? This certainly adds to the dimension of the whole thing. -- Best Regards, Shyam

2 1

Ibr-dtn v.0.10.0 ibrdtn.stats missing?
by Shyam B 10 Sep '13

10 Sep '13

Hi DTN folks/masters, I am using Raspberry PI. I just updated the Raspbian Wheezy and upgraded to the latest* version 0.10.0 of ibrdtn*. I just did (sudo apt-get ibrdtnd ibrdtn-tools). Unfortunately, I cannot see the */tmp/ibrdtn.stats* in the folder. The stats commands are also missing in the ibrdtn.conf file. I have all the required permissions for this folder. In my own program, *I read this file*to get the neighbor node's status (i.e. if they are available) and not I am unable to. This feature was available in the previous version of ibrdtn (which I am unable to retrieve now - I cannot see the earlier version in the Raspbian repository! In the version update: http://trac.ibr.cs.tu-bs.de/project-cm-2012-ibrdtn/wiki/changelog: It says: - Statistic Logger replaced by API management calls What does this mean? If there is an API, where is it and what is the easy way of calling *this API*? I am programming in python.. Also, I would like to know the difference between DTN2 and IBRDTN, which one is better to use and why/ why not? Thank you and I hope I can get some help! -- Best Regards, Shyam

3 3