How can I speed up the shutdown of gazebo with pr2_controler_manager? [closed]
When shutting down gazebo with I often get the following message:
[gazebo-2] killing on exit
[WARN] [WallTime: 1317250619.639809] [483.903000] Spawner couldn't reach pr2_controller_manager to take down controllers.
and then after about 10 seconds:
[gazebo-2] escalating to SIGTERM
Is there a way, other than just killing it with kill
, to make gazebo go down more quickly?
Closed for the following reason
question is not relevant or outdated by
tfoote
close date 2013-02-03 07:59:01
Comments
You could still maintain a clean shutdown if you type 'killall -9 gazebo' in a different terminal while the ctrl-c is waiting.
hsu
(
2011-09-28 12:03:21 -0500
)edit
It looks like a new while loops are introduced, preventing shutdown (see this [ticket](https://code.ros.org/trac/ros-pkg/ticket/4882)). Is it possible for you to create a backtrace of the hanging code?
thanks,
John
hsu
(
2011-09-28 12:25:00 -0500
)edit
I'll try to attach gdb to the relevant process when the hang happens and see what I get
Asomerville
(
2011-09-28 13:27:54 -0500
)edit
No luck trying to catch any info from gazebo with gdb. I did track down that the message is coming from pr2_mechanism/pr2_controller_manager/scripts/Spawner though I'm not familiar enough with how they interact to understand why that would make a difference.
Asomerville
(
2011-09-28 15:26:24 -0500
)edit
@Asomerville: I asked this question (http://answers.ros.org/question/2317/how-to-delay-escalation-to-sig-term-after-sending) which should help us debug this problem. thanks for trying.
hsu
(
2011-09-28 19:06:15 -0500
)edit
@hsu I chased this into gazebo and noticed that there's a "setUserQuit" method that gets called on any signal and interrupts the server only if it's headless. Otherwise the "setUserQuit" method does nothing to the wx app loop from what I can tell.
Asomerville
(
2011-10-11 13:37:56 -0500
)edit
add a comment