"Frames didn’t arrive within 5 seconds" – Intermittent Issue with RealSense Camera on Linux #13365
Comments
|
Hi @nurietarc The error Failed to read busnum/devnum is a very rare one. Usually when it has occurred, it has been related to a problem in the USB system, such as insufficient power being provided to the camera by the USB port to meet its power-draw needs. Frames didn't arrived within 5 seconds would also suggest that new camera frames have stopped arriving at the computer from the USB port. RealSense 400 Series cameras can run indefinitely so long as they remain within recommended operating temperature tolerances (not more than 35 degrees C internal temperature and not less than 0 degrees C). If a problem does occur in a long-run session (3 to 4 days before the problem occurs), it is likely to be related to the computer hardware, its operating system or the USB hardware (port / hub / controller / cable) rather than a fault with the camera hardware. If a computer has a USB power-saving plan set then I would recommend disabling that to ensure that there is not a situation where the computer may reduce USB port power and make the camera unable to operate. |
|
Thank you for your help. First, these are not long-run sessions. We are still testing in the lab, and the tests last about 10 minutes to ensure everything works well. The camera shouldn’t be overheating or anything like that. Regarding the error you mentioned, I removed any power-saving configurations on the laptop. I upgraded the kernel to 6.8 again (as mentioned, downgrading to 6.5 didn’t allow me to install the RealSense library, and since the library works fine on other machines with kernel 6.8, I gave it another try). I purged the previous installation and reinstalled the library using Option 1. Upon opening the RealSense Viewer, I got the following error: Additionally, I got this warning: Despite these errors and warning and a "frames didn't arrive" error at first, I was able to connect and view the camera in RealSense Viewer after some fiddling around, even with the errors still showing. Then I proceed to removed the redundancies as indicated by the error (although I’m not entirely sure if I did this step correctly, I may need some guidance...), and just to be sure, I purged and reinstalled again. The redundancy and busnum/devnum errors persist and now I can't vew the camera RGB signal in RealSense Viewer anymore. Below is the complete log: Thanks again! |
|
You are very welcome! Multiple udev rules (/lib and /etc) occur when librealsense has been installed from packages and from source code on the same computer. One method or the other (packages or source) should be used but not both, otherwise you could have two versions of librealsense on the same computer that conflict with each other. When the multiple udev rules error occurs, the rule that should always be removed is As the above log shows that the multiple udev rules error is still present, please try deleting the /etc rule to see whether any of the errors that follow afterwards disappear. The udev rule is responsible for handling RealSense camera devices in Ubuntu. |
|
Hello, I have removed the rule doing Thank you! |
|
The ROS launch log atthe beginning of this case indicates that the camera was being detected as being on a slow USB 2.1 connection instead of a fast USB 3.1 or 3.2 connection.
A camera may be mis-detected as being on a USB2 connection even if it is plugged into a USB3 port. It may work fine on one machine but have problems with another. This may be due to the USB controller that a particular computer manufacturer has chosen for handling devices plugged into the USB ports, as some USB controllers will work with RealSense cameras better than others. Testing the camera with it attached to a USB hub rather than plugged directly into the computer is the best way to test whether there is a controller compatibility issue, as a USB hub will use its own brand of USB controller different from the controller that the computer's built-in USB ports are using. In a case at the link below, a RealSense user found that their busnum/devnum problem disappeared when the camera was used on a different computer (a laptop). https://support.intelrealsense.com/hc/en-us/community/posts/27878881494419/comments/27902533368851 |
|
Hi Marty, I am aware of the Since you have said that the laptop itself and its connectors could be the problem behind, I have tried to open the RealSense viewer in another laptop which has exactly the same specifications (ThinkPad L390). I get the Frames didn't arrive error but no the busnum/devnum. Here is the log: I am even more lost than before... haha Thank yoy again very much! |
|
As I was telling you, sometimes these errors seem to appear "suddenly." Let me explain what happened now: I have noticed that in the log of the previous comment (link here), there were no warnings related to the USB 2.1 port, despite the RealSense Viewer indicating it was connected via USB 2.1 in the top left corner. Therefore, I wanted to see if making small adjustments would trigger that error in the log. Simply starting and stopping the RGB profile streaming caused the camera to begin functioning properly, and frames started arriving without any further intervention (in the same computer that 5 minutes, reported the log of my last comment!). Taking advantage of the camera's functionality, I have adjusted some settings (like battery saving options and whether the laptop was plugged in or not), and I also have connected and disconnected the camera to see if that would cause it to stop working. The response is no... it continues to function without issues. Here is the log: This raises the same question: why (and how to avoid that) do these errors occur sporadically, sometimes on the same computer and at other times on different machines, even under similar conditions? |
|
It is difficult to mitigate against random events. You might be able to reduce the likelihood of them occurring though if you build the librealsense SDK from source code with the libuvc backend or the RSUSB backend installation methods though (if you are not using them already). A build of librealsense created with these methods bypasses the Linux kernel and so can avoid unpredictable errors caused by conflicts between librealsense and the kernel. |
|
I'll try to build the librealsense with RSUSB backend once again, thank you. However, would this solve the "frames didn't arrive" error? Isn't it a power-related error? Thank you! |
|
It would not typically be a power-related error. It could occur because of causes such as:
|
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Context
I am using the Intel RealSense camera for robotics applications and trying to implement the ROS Wrapper for Intel (link here) but having issues getting it to work.
Steps followed and errors (and how to reproduce them)
Attempt 1:
Attempt 2:
dpkg -l | grep "realsense" | cut -d " " -f 3 | xargs sudo dpkg --purge) and reinstalled the SDK 2.0 library using Option 1 (Option 1: Install librealsense2 debian package from Intel servers), following the steps described here.Attempt 3:
dpkg -l | grep "realsense" | cut -d " " -f 3 | xargs sudo dpkg --purge) and attempted to reinstall the SDK, but encountered a new kernel error during installationTests on other machines
Prior attempts to resolve the issue
initial_resetparameter.The problem
Even though the camera works on some machines, we cannot pinpoint the root cause of the issue or why the error occurs intermittently between different computers, and even on the same machines, where it may "magically" stop working after a few days without any apparent reason. This makes the camera unreliable for robotics applications, as we never know when the next unexpected error might appear. Any suggestions?
Many thanks!
The text was updated successfully, but these errors were encountered: