Robotics StackExchange | Archived questions

Gazebo is aborting

Hello,

I am currently trying to launch a simulation of TIAGo and Gazebo is aborting just a few seconds after the process has started running. In first place I thought that it was due a weird error related with executing Python nodes, the file had no execution permissions and ROS could not find the node. I mend that just executing

chmod +x python_file.py

but the error of Gazebo aborting persisted. I don't understand why but I think it's fully related with ROS Python packages. Here is the error of the roslaunch:

[gazebo_gui-3] process has died [pid 57438, exit code 134, cmd /.../gazebo_ros/gzclient __name:=gazebo_gui __log:=~/.ros/log/4602d00c-8129-11e8-9ca3-000c2954f4ae/gazebo_gui-3.log].
log file: ~/.ros/log/4602d00c-8129-11e8-9ca3-000c2954f4ae/gazebo_gui-3*.log

[tuck_arm-27] process has finished cleanly
log file: ~/.ros/log/4602d00c-8129-11e8-9ca3-000c2954f4ae/tuck_arm-27*.log

I am not sure if the reason of the error is in the log specified by Gazebo a few lines after the gazebo_gui-3 process died but in the log I found the following:

[rospy.internal][WARNING] 2018-07-06 14:31:44,659: Unknown error initiating TCP/IP socket to development:47125 (http://development:46830/): Traceback (most recent call last):
  File "/opt/ros/kinetic/lib/python2.7/dist-packages/rospy/impl/tcpros_base.py", line 557, in connect
    self.read_header()
  File "/opt/ros/kinetic/lib/python2.7/dist-packages/rospy/impl/tcpros_base.py", line 650, in read_header
    self._validate_header(read_ros_handshake_header(sock, self.read_buff, self.protocol.buff_size))
AttributeError: 'NoneType' object has no attribute 'buff_size'


[rospy.internal][ERROR] 2018-07-06 14:32:04,856: Traceback (most recent call last):
  File "/opt/ros/kinetic/lib/python2.7/dist-packages/rospy/impl/tcpros_base.py", line 729, in receive_once
    self.stat_bytes += recv_buff(sock, b, p.buff_size)
  File "/opt/ros/kinetic/lib/python2.7/dist-packages/rospy/impl/tcpros_base.py", line 103, in recv_buff
    d = sock.recv(buff_size)
error: [Errno 104] Connection reset by peer

[rospy.internal][WARNING] 2018-07-06 14:32:04,857: Unknown error initiating TCP/IP socket to development:43093 (http://development:32888/): Traceback (most recent call last):
  File "/opt/ros/kinetic/lib/python2.7/dist-packages/rospy/impl/tcpros_base.py", line 558, in connect
    self.local_endpoint = self.socket.getsockname()
AttributeError: 'NoneType' object has no attribute 'getsockname'

I don't know what could be... Do you have any hint of what could be?

Thank you, Marc

Asked by mdominguez on 2018-07-06 10:33:40 UTC

Comments

I am also facing a similar issue with ros-kinetic and px4 running on Ubuntu 16. Does anyone have update on this. One thing i did try is that this issue does not arise with ros-melodic and px4 on ubuntu 18.

Asked by argos on 2020-03-26 23:26:03 UTC

@mdominguez @argos , have either of you found a solution to this issue? I am having the same problem as is stated in the OP.

Asked by zaneytop on 2020-11-29 11:44:20 UTC

Answers

From the log it appears that your system is refusing to open sockets. https://answers.ros.org/question/11488/unknown-error-initiating-tcpip-socket/

I found reference that you might be exhausting your socket allocations: https://answers.ros.org/question/202551/nodes-stop-subscribing-over-time/

Or possibly some other system setting is preventing them from opening such as the packet size: https://answers.ros.org/question/10248/prosilica-intrinsincs-parameters-failed-to-load/

Asked by tfoote on 2018-07-09 02:59:43 UTC

Comments