Shadow interface crashes after short while

Bug #872835 reported by ErikBerglund on 2011-10-12
This bug affects 1 person
Affects Status Importance Assigned to Milestone
sr ROS interface

Bug Description

A short while after starting the shadow interface the /srh/ or /sr_arm interfaces stop publishing updates.

The time varies from under 20 seconds up to 10 minutes.

Once the bug appears, no more updates are published to:

The bug appears equally likely to disable the /srh as the /sr_arm interfaces, sometimes both at the same time.

After the updates stop, sending commands to /sendupdate has no effect.

Full logs attached.

Related branches

Ugo (ugocupcic) wrote :

Hi Erik, thanks for reporting the bug.
Can you tell me a bit more about it?
 - which version are you using?
 - which command do you use to start the interface?

Changed in sr-ros-interface:
assignee: nobody → Ugo (ugocupcic)

Version: Diamondback, bazaar revision 233.

I start gazebo and the interface like this (in separate consoles):
roslaunch sr_hand gazebo_arm_and_hand_motor.launch
roslaunch sr_hand sr_arm_motor.launch

Ugo (ugocupcic) wrote :

Same as bug 844934:

Gazebo crashes with this version on the trunk :

To solve this problem, you need to comment out publishMechanismstatistics() in pr2_controller_manager/src/controller_manager.cpp and recompile the pr2_controller_manager... I'm still waiting for an answer on why this is not working :S

Hope this helps.

Tried solution proposed in #4, but the problem persists.

Am I recompiling pr2_controller correctly?

I did this:
roscd pr2_controller_manager
sudo chmod o+w . -R
<edit src/controller_manager.cpp>
rosmake --rosdep-install pr2_controller_manager

Did I miss anything obvious?

This bug appears unrelated to bug 844934, as Gazebo starts and runs fine, but no status updates are published by the Shadow interface.

Ugo (ugocupcic) wrote :


GAZEBO=1 rosmake sr_hand sr_mechanism_model pr2_controller_manager gazebo

This bug has only been encountered in the manipulation branch.

By slowing down the amount of joint_data objects sent to the interface to about 8 per second the bug becomes rarer, but does not go away completely.

Ugo (ugocupcic) wrote :

Probably a bad mutex somewhere.

Changed in sr-ros-interface:
status: New → Confirmed
Ugo (ugocupcic) on 2011-12-14
Changed in sr-ros-interface:
status: Confirmed → Fix Committed
To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers

Bug attachments