rostopic groovy strangeness [closed]
Hello, I'm not sure if this is an issue with rostopic or with the hector_quadrotor, but I recently upgraded to groovy and when testing the current state of the quadrotor stack today I noticed the following occasional "glitch" in the reported rates:
$ rostopic hz /front_cam/camera/image
subscribed to [/front_cam/camera/image]
average rate: 10.204
min: 0.069s max: 0.127s std dev: 0.01776s window: 9
average rate: 10.045
min: 0.069s max: 0.127s std dev: 0.01631s window: 19
average rate: 9.954
min: 0.069s max: 0.127s std dev: 0.01572s window: 29
average rate: 9.965
min: 0.069s max: 0.127s std dev: 0.01434s window: 38
average rate: 9.977
min: 0.069s max: 0.127s std dev: 0.01435s window: 48
average rate: 9.944
min: 0.069s max: 0.127s std dev: 0.01385s window: 58
average rate: 9.951
min: 0.069s max: 0.127s std dev: 0.01330s window: 68
average rate: 9.957
min: 0.069s max: 0.127s std dev: 0.01265s window: 78
average rate: 9.942
min: -16061.683s max: 16061.895s std dev: 2435.28216s window: 88
average rate: 9.961
min: -16061.683s max: 16061.895s std dev: 2318.31955s window: 97
average rate: 9.955
min: -16061.683s max: 16061.895s std dev: 2206.25640s window: 107
average rate: 9.956
min: -16061.683s max: 16061.895s std dev: 2109.01622s window: 117
average rate: 9.966
min: -16061.683s max: 16061.895s std dev: 2023.59521s window: 127
average rate: 9.939
min: -16061.683s max: 16061.895s std dev: 1947.77802s window: 137
average rate: 9.951
min: -16061.683s max: 16061.895s std dev: 1879.89021s window: 147
The values change when rostopic is re-run and the onset of the behavior also varies E.G.
min: -15800.895s max: 15801.091s std dev: 2382.08932s window: 89
min: -15766.858s max: 15767.068s std dev: 2145.61190s window: 109
min: -15748.591s max: 15748.802s std dev: 5109.55185s window: 20
It also occurs for other topics. Any suggestions on where to begin checking for the source of the problem? This seems like it may be a hard one to track down as it only sometimes occurs.
All the topics where this was observed are published by the node /gazebo.
Are your header timestamps being generated from multiple boxes without synchronized times?
All nodes are running on the same machine, and using rxplot and rostopic echo it seems
header/stamp/secsandheader/stamp/nsecsare changing as one would expect for each topic...Are all the affected topics being published by a common source node?
gazebo... hmm...
Yep. Sounds like bug to me, too.
rebooted the machine.. and now haven't been able to observe the errant behavior.. (the machine had been rebooted twice previously since groovy was installed, but gazebo had not been tested previously)