Hello Sergey,

this looks like a serious bug. To get a hint where the fault happens, we need to investigate the stack trace of the crash. For that, it is necessary to install the debug symbols and the gdb with this command:

sudo apt-get install ibrdtnd-dbg ibrdtn1-dbg ibrcommon1-dbg gdb

Then you need to run the test again. But before you start the daemon you need to raise the ulimit for core dumps before running each daemon.

ulimit -c unlimited

If the daemon crashs again it, there should be a core file (named core or core.xxx). If there is that file, you can examine the stack trace with gdb like this... (replace <corefile> with the available core file).

gdb --core <corefile> dtnd

(gdb) set pagination off (gdb) set logging on (gdb) bt ... (gdb) t a a bt ... (gdb) set logging off (gdb) quit

After that there exists a gdb.txt file in your directory and it would be very nice, if you could send that file to me. It will help me to localize the issue.

Kind regards, Johannes Morgenroth

Am 19.08.2013 11:16, schrieb ssireskin@gmail.com:

Hi,

I have set up several ibr-dtn nodes with multiple interconnections, without discovery and with static routing. I have defined all possible routes in ibrdnd config files to test path fail-over. When I start dtnping, dtnd crashes after several successful pings. Then ping continues via an alternative route, until the other intermediate node crashes. Sometimes the end node, which is being pinged, crashes first.

I have attached a network diagram, ibrdtnd config files and logs.

I was pinging node 226 from node 224. The route was 224 -> 225 -> 226. The node 226 has crashed. The route changed to 224 -> 227 -> 226. The node 227 has crashed.

Please, advise, what could be done to fix this issue.

-- Best regards, Sergey Sireskin