Brian Gerkey さんは書きました: > On Thu, Mar 4, 2010 at 3:51 AM, Klaus Petersen > wrote: >> I have a very basic problem: First, I establish a connection between a >> publisher node and a subscriber node. At this point I can receive >> messages from the publisher with the subscriber node. But when I >> re-start my publisher node (quit and run again) the subscriber does not >> automatically resubscribe to the message. I have to restart the >> subscriber as well to be able to receive again. Is this the normal >> behavior with ros or am I doing something wrong? I have quite a big >> robot control program tied to one of the nodes, so restarting along the >> way (the initialization of the robot takes ages) is very troublesome. > > hi Klaus, > > Something's going wrong there. When your publisher is restarted, it > should re-advertise with the master, which should send the new > publisher list to your subscriber, which should then re-connect to > your publisher. > > You can test this behavior with roscpp_tutorials/talker and > roscpp_tutorials/listener; no matter in what order they're > (re)started, listener will always connect to talker. > > Can you tell us something about your publisher and subscriber? Is > there anything unusual about your subscriber? > > brian. > _______________________________________________ > ros-users mailing list > ros-users@code.ros.org > https://code.ros.org/mailman/listinfo/ros-users > Hi Brian, The problem is consistent for the tutorials nodes. I think there is someting wrong with the communication of my nodes with the master. My system is running on one linux, one macos and one windows box. I had trouble running roscore on the linux computer for some reason. After running it on the mac everything seemed to be ok until I encountered the reported problem. I think it comes down to some trouble with my linux box. I did the connectivity checks from ros.org, which showed some problems with name resolution in my network. I fixed that by setting ros_ip. Still it seems like some ports on my ubuntu box let go data in, but no data go out. Is there any common mistake that leads to that symptom? Anyway, I guess general network trouble is out of the scope of this mailing list :) Klaus