Cedric and Joel, It turns out that Josh was right, rospy is writing all debug statements to log files, even if the log level is not set to debug. I've filed a ticket against it, which when addressed, should fix these issues. You can find the ticket here: https://code.ros.org/trac/ros/ticket/2916 As far as rxgraph goes, I'm not sure whether that level of logging is needed anymore. I have a feeling that its leftover from when the tool was much less stable and we frequently had to debug it. I've filed another ticket for this here: https://code.ros.org/trac/ros/ticket/2918 Hope all is well, Eitan On Mon, Jul 26, 2010 at 10:51 AM, Joel Cartwright wrote: > Eitan, > > On the subject of logging, I noticed today that my ~/.ros/log directory > had grown quite large, and rxgraph was producing particularly verbose > logs. DEBUG logging seems to be enabled at line 68 of > cturtle/ros/tools/rxgraph/src/rxgraph/__init__.py. Is this level of > logging intentional? > > Regards, > > Joel > > Eitan Marder-Eppstein wrote, On 26/07/10 18:32: > > Cedric, > > > > We normally only enable the DEBUG log level when we're actively testing > > or trying to diagnose a problem in a node. Is there any reason that you > > need to have the log level for actionlib set to DEBUG? Can you try > > setting it to INFO instead to see if that solves your load/logfile > > problems. If it doesn't, please let us know as that's a bug for the ROS > > logging system. > > > > Hope all is well, > > > > Eitan > > > > On Sat, Jul 24, 2010 at 3:54 AM, Cedric Pradalier > > > > > wrote: > > > > Hi, > > > > I've just noticed two rospy.logdebug statement in cturtle > > action_server.py and simple_action_server.py: > > In action_server.py:314: rospy.logdebug("Status async"); > > In simple_action_server.py:306: rospy.logdebug("SAS: execute"); > > > > On my system, with my simple launch files and simple_action_server, > > these two are making my logfiles grow to infinity, and makes my > > action_server use 10% of my CPU while not doing anything. > > I've commented them out for now, but is there a smarter way? > > > > For information, I've also noticed that rosout is in general using > > around 10% of my CPU with a slow harddrive (within virtualbox), but > that > > it goes down to 4-5% when putting the ROS_LOG_DIR in tmpfs. This > might > > be significant for embedded system logging on a flash-drive. > > > > -- > > Dr. Cedric Pradalier > > http://www.asl.ethz.ch/people/cedricp > > > > _______________________________________________ > > ros-users mailing list > > ros-users@code.ros.org > > https://code.ros.org/mailman/listinfo/ros-users > > > > > > > > > > _______________________________________________ > > ros-users mailing list > > ros-users@code.ros.org > > https://code.ros.org/mailman/listinfo/ros-users > > -- > Research Associate > Ocean Systems Laboratory > Heriot-Watt University, UK > _______________________________________________ > ros-users mailing list > ros-users@code.ros.org > https://code.ros.org/mailman/listinfo/ros-users >