-
Notifications
You must be signed in to change notification settings - Fork 486
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
gzclient sometimes empty on start #681
Comments
Original comment by Brian Gerkey (Bitbucket: Brian Gerkey, GitHub: gerkey).
|
Original comment by Steve Peters (Bitbucket: Steven Peters, GitHub: scpeters). I just ran into a predictable way to reproduce this and have been debugging it with @nkoenig A partial fix is the following:
|
Original comment by Steve Peters (Bitbucket: Steven Peters, GitHub: scpeters). partial fix in pull request #815 |
Original comment by Steve Peters (Bitbucket: Steven Peters, GitHub: scpeters). A partial fix was merged in pull request #815; I'm leaving this open however, until we have some tests for request / response messages. |
Original comment by Steve Peters (Bitbucket: Steven Peters, GitHub: scpeters). May be related to #765 |
Original comment by Steve Peters (Bitbucket: Steven Peters, GitHub: scpeters). Issue #1234 was marked as a duplicate of this issue. |
Original comment by Steve Peters (Bitbucket: Steven Peters, GitHub: scpeters). I reported a similar issue in #2046 related to loading a heightmap visual without a collision. As a hint, try attaching gdb to any hanging gzclient processes and see if it is blocking on a |
Original comment by Nate Koenig (Bitbucket: Nathan Koenig).
|
Original comment by Nate Koenig (Bitbucket: Nathan Koenig).
|
potential fix for gazebo11 in #3121 |
it should be improved now since #3121 was merged, so I'm going to mark this as closed please post here with details if the problem recurs |
I received the upgrade to 11.9.0 today. This version crashes one of my simulations. I narrowed the problem down to this commit 720b52d The crash happens after I see this message
and the stack trace is the following
I had the same problem on another occasion which you can check out here nlamprian/gazebo_video_monitors@5eebfa6 I have no clue what's going on with the scene, and the underlying problem is likely not directly related to #3121, but please consider reverting and releasing these changes. It would be hugely appreciated. |
can you provide the backtrace for other threads as well with
given that #3121 was merged after nlamprian/gazebo_video_monitors@5eebfa6, I don't think it caused the failure in
I'm willing to consider reverting #3121, but I need to understand more first. Do you have a way to reproduce the first failure that you posted? |
I only mentioned this to highlight that there is a bug somewhere else that gets triggered by your changes.
Unfortunately, the only way I know how to reproduce the problem is connected to one of the models I'm working with, and I cannot share that publicly. |
can you share the number and types of sensors and plugins in your model? |
There are 42 ray sensors, 1 imu, 1 camera, 4 depth cameras, 17 model plugins, 18 sensor plugins, 4 visual plugins, and 4 world plugins. |
these are the threads in the backtrace that jump out at me:
It looks like some ray sensors are already running before all the meshes are loaded. Do you have multiple models with ray sensors or are they all in a single model? 42 ray sensors is quite a significant number; I don't think our current regression tests have anything close to that |
All sensors are in one model. I'm using sets of ray sensors to simulate light curtains (hence the smaller number of sensor plugins). You are right to assume this is related to mesh files. I have two (similar) models, which are two different versions of the same product. The problem occurs with the first model, but not the second one. In the second model, I have simplified collision geometries, and I guess that makes a difference. Also, I tested deleting all the collisions from the first model. Gazebo starts normally in that case. If you decide to go after this bug, let me know if I can help (I would ask for some context before I start digging). In the meantime, a temporary workaround would really help me a lot here. |
Original report (archived issue) by Brian Gerkey (Bitbucket: Brian Gerkey, GitHub: gerkey).
Occasionally, when running
gazebo
, thegzclient
window comes up empty. In this situation, starting anothergzclient
works fine. Seems like a startup race condition.The text was updated successfully, but these errors were encountered: