IntelRealSense
Sometimes depth stream blocked on single-frame pipe
| Required Info | |
|---|---|
| Camera Model | { D400 } |
| Operating System & Version | {Ubuntu 20} |
Issue Description
A simple code of this kind sometimes (perhaps after hundreds or thousands of iterations) blocks the depth stream and prevents new frames from being received
//config enabled only depth stream
pipe->start(*config);
pipe->wait_for_frames();
pipe->stop();
In realsense viewer, if you enable the display of depth frames, you will see the message "No frames recieved" The only thing that helps is the physical plugging of the wire. I checked on several different sensors and wires. Is there any way to avoid this behavior?
Hi @reuksv How does it behave if you use a frames variable with wait_for_frames() and then use the keep() instruction to store the frames in the computer's memory?
//config enabled only depth stream
pipe->start(*config);
frames = pipe->wait_for_frames();
frames->keep();
pipe->stop();
Hi @reuksv How does it behave if you use a frames variable with wait_for_frames() and then use the keep() instruction to store the frames in the computer's memory?
//config enabled only depth stream pipe->start(*config); frames = pipe->wait_for_frames(); frames->keep(); pipe->stop();
Hello @MartyG-RealSense No changes have occurred.
I would next recommend using an Ubuntu system monitoring tool such as htop to monitor the available memory on the computer whilst the program is running and see whether it is progressively declining over time as the program repeatedly iterates. Such a reduction is known as a memory leak and can result in a program becoming unstable or crashing once the free memory becomes low enough.
I would next recommend using an Ubuntu system monitoring tool such as htop to monitor the available memory on the computer whilst the program is running and see whether it is progressively declining over time as the program repeatedly iterates. Such a reduction is known as a memory leak and can result in a program becoming unstable or crashing once the free memory becomes low enough.
@MartyG-RealSense I've been monitoring, there are no leaks. I also clarified that the Fail happens very randomly. This can happen on the tenth iteration, for example.
Does the issue still occur if you put a value in the brackets of the wait_for_frames instruction such as 1000?
pipe->wait_for_frames(1000);
How does the program behave if the stop instruction is commented out and the pipeline is permitted to run continuously without multiple start-stops?
If your project requires repeated start-stop cycles instead of running continuously then you could try putting a short sleep command before the pipe start line to see whether a small time wait before starting the pipeline also helps to reduce the likelihood of a problem occurring.
Next, please try placing a camera reset instruction before the pipe start line in your script using the Python code below. This should have the same effect as performing an unplug-replug of the USB cable.
ctx = rs.context()
devices = ctx.query_devices()
for dev in devices:
dev.hardware_reset()
Hi @reuksv Do you require further assistance with this case, please? Thanks!
@MartyG-RealSense Unfortunately, I haven't had time to check the last tip yet. Please do not close the ticket yet
It's no problem at all to keep it open. Thanks very much for the update!
Hi @reuksv Do you require further assistance with this case, please? Thanks!
Thank you. I think hardware reset should help, but I managed to test it only in the realsense viewer interface. You can close the ticket.
As you suggested, I will close the issue. Thanks very much for the update!