[ros-users] slam_gmapping crashing

Gonçalo Cabrita goncabrita at gmail.com
Wed Jul 28 15:18:35 UTC 2010


Hi Joel, thanks for your reply.

I haven't tested it again on my computer. We installed ROS on another
laptop, also a Core2Duo but with more RAM (4Gb) and it is working perfect
now.

I suppose it could be a memory issue since it seems slam_gmapping is taking
over 800Mb of RAM and I only have 1Gb, I'll get back to you once I try
running slam_gmapping again on my MacBook.

Thanks for all your help,

Gonçalo Cabrita
ISR - University of Coimbra
Portugal

On Wed, Jul 28, 2010 at 10:48 AM, Joel Cartwright
<J.J.Cartwright at hw.ac.uk>wrote:

> Gonçalo,
>
> Have you checked your /var/log/messages for lines like the following?
>
> Out of Memory: Killed process [PID] [process name].
>
> If you have lines like those, then it is the Linux kernel that is
> killing your process with SIGKILL.
>
> Joel
>
> Gonçalo Cabrita wrote, On 28/07/10 00:38:
> > Hi Brian,
> >
> > I tried gdb and valgrind and both tell me that slam_gmapping is getting a
> SIGKILL and that's it.
> >
> > Tomorrow in the morning I'll install Ubuntu and ROS on my home Mac to see
> what happens.
> >
> > I'm running the same Ubuntu on my Mac that we have on the Eee PCs with
> the same version of ROS installed. The Eee PCs are able to run slam_gmapping
> without ever crashing (not so bad after all, although the maps do come out a
> bit weird).
> >
> > I'll let you know of my progress!
> >
> > Gonçalo Cabrita
> > ISR - University of Coimbra
> > Portugal
> >
> > On Jul 27, 2010, at 8:20 PM, Brian Gerkey wrote:
> >
> >> hi Gonçalo,
> >>
> >> It's odd that slam_gmapping is receiving a SIGKILL.  Can you try
> >> running it in gdb, outside roslaunch?  Something like:
> >>
> >> $ roscd gmapping
> >> $ gdb bin/slam_gmapping
> >> ...
> >> (gdb) r
> >>
> >> If it still dies from SIGKILL, then there's nothing that gdb is going
> >> to be able to tell you.  If it dies from something else, then you
> >> should try to get a backtrace from gdb.
> >>
> >> You might try running inside valgrind, to see if you get a different
> result:
> >>
> >> $ roscd gmapping
> >> $ valgrind -v bin/slam_gmapping
> >>
> >>      brian.
> >>
> >> 2010/7/27 Gonçalo Cabrita<goncabrita at gmail.com>:
> >>> Thanks for your quick reply Brian.
> >>> slam_gmapping is crashing randomly, it can crash a few seconds or a few
> >>> minutes after it started. Also I've noticed that slam_gmapping crashing
> is
> >>> causing the internet connection to crash as well.
> >>>
> >>> I did what you suggested and this is what I got:
> >>> I did a launch file as instructed:
> >>> <launch>
> >>>    <master auto="start"/>
> >>>    <node pkg="gmapping" type="slam_gmapping" respawn="false"
> >>> name="slam_gmapping" output="screen" launch-prefix="xterm -e gdb
> --args">
> >>>    </node>
> >>> </launch>
> >>> I open a roscore, set use_sim_time to true and then I launch
> slam_gmapping:
> >>> $ roslaunch slam_gmapping.launch
> >>> I get the gdb xterm window...
> >>> (gdb) set logging on
> >>> (gdb) run
> >>> Now I start my rosbag play....
> >>> The output in the gdb.txt file is:
> >>> Starting program:
> >>> /home/blazz/ros/stacks/slam_gmapping/gmapping/bin/slam_gmapping>
>  gdb_output
> >>> [Thread debugging using libthread_db enabled]
> >>> [New Thread 0xb7fdcb70 (LWP 2249)]
> >>> [New Thread 0xb77dbb70 (LWP 2250)]
> >>> [New Thread 0xb6fdab70 (LWP 2251)]
> >>> [New Thread 0xb65ffb70 (LWP 2258)]
> >>> [New Thread 0xb5dfeb70 (LWP 2261)]
> >>> [New Thread 0xb55fdb70 (LWP 2304)]
> >>> [New Thread 0xb4dfcb70 (LWP 2305)]
> >>> [Thread 0xb4dfcb70 (LWP 2305) exited]
> >>> [Thread 0xb55fdb70 (LWP 2304) exited]
> >>> [Thread 0xb5dfeb70 (LWP 2261) exited]
> >>> [Thread 0xb65ffb70 (LWP 2258) exited]
> >>> [Thread 0xb6fdab70 (LWP 2251) exited]
> >>> [Thread 0xb77dbb70 (LWP 2250) exited]
> >>> [Thread 0xb7fdcb70 (LWP 2249) exited]
> >>> Program terminated with signal SIGKILL, Killed.
> >>> The program no longer exists.
> >>> The output doesn't give too much info... Does it help?
> >>> Gonçalo Cabrita
> >>> ISR - University of Coimbra
> >>> Portugal
> >>> On Tue, Jul 27, 2010 at 4:40 PM, Brian Gerkey<gerkey at willowgarage.com>
> >>> wrote:
> >>>>
> >>>> hi Gonçalo,
> >>>>
> >>>> A shot in the dark: you might be exhausting available memory.  In any
> >>>> case, try running slam_gmapping in gdb to see what's causing it to die
> >>>>
> >>>> (
> http://www.ros.org/wiki/roslaunch/Tutorials/Roslaunch%20Nodes%20in%20Valgrind%20or%20GDB
> ).
> >>>>
> >>>>         brian.
> >>>>
> >>>> 2010/7/27 Gonçalo Cabrita<goncabrita at gmail.com>:
> >>>>> Hi everyone!
> >>>>> I'm trying to generate a map according to the tutorials and I've ran
> >>>>> into a
> >>>>> bit of a problem. I'm trying to run slam_gmapping on my MacBook
> Core2Duo
> >>>>> running Ubuntu 10.04 with ROS and I'm allways getting the same error,
> >>>>> slam_gmapping crashes after a while after starting to run...
> >>>>> ...
> >>>>> *************RESAMPLE***************
> >>>>> Deleting Nodes: 3 4 5 7 10 12 17 19 21 24 26 27 Done
> >>>>> Deleting old particles...Done
> >>>>> Copying Particles and  Registering  scans... Done
> >>>>> update frame 422
> >>>>> update ld=1.00063 ad=0.0320879
> >>>>> Laser Pose= 5.26866 0.0983503 0.0260777
> >>>>> m_count 4
> >>>>> Average Scan Matching Score=110.945
> >>>>> neff= 29.3825
> >>>>> Registering Scans:Done
> >>>>> /home/blazz/ros/ros/bin/rosrun: line 35:  1862 Killed
> >>>>>   $exepath "$@"
> >>>>> This is running on other machines in the lab but not in my Mac. I've
> >>>>> tried
> >>>>> re-installing ROS but the problem persists.
> >>>>> What additional information should I provide to help solve this
> >>>>> problem???
> >>>>> Thanks for your help,
> >>>>> Gonçalo Cabrita
> >>>>> ISR - University of Coimbra
> >>>>> Portugal
> >>>>> _______________________________________________
> >>>>> ros-users mailing list
> >>>>> ros-users at code.ros.org
> >>>>> https://code.ros.org/mailman/listinfo/ros-users
> >>>>>
> >>>>>
> >>>> _______________________________________________
> >>>> ros-users mailing list
> >>>> ros-users at code.ros.org
> >>>> https://code.ros.org/mailman/listinfo/ros-users
> >>>
> >>>
> >>> _______________________________________________
> >>> ros-users mailing list
> >>> ros-users at code.ros.org
> >>> https://code.ros.org/mailman/listinfo/ros-users
> >>>
> >>>
> >> _______________________________________________
> >> ros-users mailing list
> >> ros-users at code.ros.org
> >> https://code.ros.org/mailman/listinfo/ros-users
> >
> > _______________________________________________
> > ros-users mailing list
> > ros-users at 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 at code.ros.org
> https://code.ros.org/mailman/listinfo/ros-users
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ros.org/pipermail/ros-users/attachments/20100728/2146388b/attachment-0003.html>


More information about the ros-users mailing list