You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The number of sockets keeps steadily climbing until the open file limit is hit. rosmaster then becomes unresponsive (e.g. rostopic list or rosnode list both just hang) and starts consuming 100% CPU.
Background:
While starting rosout in a loop isn't the most sensible thing, it's just the minimal test case to reproduce this. Trying "rosnode list" or "rostopic list" doesn't leak sockets in the same way.
I've had a few instances of this happening and it probably hasn't been rosout (but due to other pressures I haven't had the time to figure out the flaky component)
The text was updated successfully, but these errors were encountered:
The master is not garbaging ServerProxy instances in the case when it notifies nodes to shut down. You likely restarted a non-anonymous node several times which leaked connections for every shut down node.
Steps to reproduce:
roscore
while :; do rosrun rosout rosout; done
lsof | grep rosmaster | grep -c CLOSE_WAIT
The number of sockets keeps steadily climbing until the open file limit is hit. rosmaster then becomes unresponsive (e.g. rostopic list or rosnode list both just hang) and starts consuming 100% CPU.
Versions:
Ubuntu 12.04.5 LTS
ros-hydro-desktop-full 1.0.2-0precise-20140807-2038-+0000
Background:
While starting rosout in a loop isn't the most sensible thing, it's just the minimal test case to reproduce this. Trying "rosnode list" or "rostopic list" doesn't leak sockets in the same way.
I've had a few instances of this happening and it probably hasn't been rosout (but due to other pressures I haven't had the time to figure out the flaky component)
The text was updated successfully, but these errors were encountered: