Closed 7oud closed 4 years ago
Hi @7oud The Xavier NX is new enough that at this point, the only Librealsense Xavier installation guides available that I know of are for the Xavier AGX model.
The JetsonHacks website has a guide for doing a simple install with a RealSense Debian repository that is compatible with Jetson AGX Xavier and does not involve patching of modules and kernels.
https://github.com/jetsonhacks/installRealSenseSDK
It may be worth trying this with your NX first, as because it is a simple process, you will not have lost much time if it does not work for NX.
JetsonHacks also have an earlier AGX Xavier installation guide involving patching that is more complex.
@MartyG-RealSense Thanks! The installation from Debian in JetsonHacks repo is the same as here. So only two methods can be tried, building from source and earlier AGX Xavier installation guide
Thank you very much @7oud - please let me know whether either of these two methods works for you. Good luck!
@MartyG-RealSense I have test the 1st method, just build the latest librealsense of master branch, as installRealSenseSDK described. However it didn't work. Most of time realsense view could not show the images. When plug out then plug in D415, viewer will show the image, if stop the stream and reopen it, "No frame received". Even the viewer show images after the 1st time plug in the camera, some error info was given "get_device_time_ms() took too long (more then 2 mSecs)". Sometimes when closed the viewer, terminal prompted "double free or corruption (out)" Could you tell the difference between the "installRealSenseSDK" and buildLibrealsense2Xavier, the latter method will be better for Xavier series?
If you have built the latest Librealsense branch (currently 2.34.0) then the error get_device_time_ms() took too long (more then 2 mSecs) is not likely related to the NX. This is a problem that has been occurring in the RealSense Viewer on Windows, Ubuntu and MacOS with various users' installation of 2.34.0, including my own. So for now I would recommend disregarding that issue as it likely has no relation to your problems with NX specifically.
The error "double free or corruption (out)" that you experienced when closing the Viewer is apparently caused during freeing up memory, which makes sense as applications try to free up memory for re-use when they are being closed.
The problems in this case have similarities to a case from November 2019 that involved Jetson.
@MartyG-RealSense yes, there are some similarities to #5315
crw-rw-rw-+ 1 root plugdev 81, 0 5月 19 17:16 video0
crw-rw-rw-+ 1 root plugdev 81, 1 5月 19 17:16 video1
crw-rw-rw-+ 1 root plugdev 81, 2 5月 19 17:16 video2
ls: cannot access 'vid*': No such file or directory
zxxx adm cdrom sudo audio dip video plugdev i2c lpadmin gdm sambashare weston-launch trusty crypto gpio
When you stop streams and restart them, are you stopping and starting the same stream or alternating between the streams please (e.g depth on RGB off, or RGB on and depth off). Sometimes the streams in the Viewer can get confused when this is done. Here is an example case:
https://github.com/IntelRealSense/librealsense/issues/2057
Do you experience the "No frames received!" error if you start the Viewer and only toggle one stream (the same stream) on and off please?
I wonder if the pipeline is remaining open when the stream is stopped the first time, preventing the pipeline from being restarted (as its status has to be that it is stopped before it can be started). Unplugging and replugging the camera would reset it, and the pipeline along with it.
Could you give some c++ code to verify your assumption, I will test it lately. And whether reopening an unclosed stream, will video devices shut down(gone)?
Approaches to C++ scripting for checking if the pipeline is active are discussed in the link below with script examples.
@7oud I have met exactly the same problem described above except I'm using D435i. I also use the https://github.com/jetsonhacks/installRealSenseSDK/buildLibrealsense.sh to install the latest librealsense. The realsense-viewer only works for the first time after the camera is connected. If I turn off the stream and turn on the same stream again, I got "No frame received".
@MartyG-RealSense I tested callback sample with some log.
rs2::log_to_console(RS2_LOG_SEVERITY_INFO);
std::map<int, int> counters;
std::map<int, std::string> stream_names;
std::mutex mutex;
auto callback = [&](const rs2::frame& frame)
{
std::cout << "come into callback... " << std::endl;
std::lock_guard<std::mutex> lock(mutex);
if (rs2::frameset fs = frame.as<rs2::frameset>())
{
for (const rs2::frame& f : fs)
{
counters[f.get_profile().unique_id()]++;
std::cout << "counters = " << counters[f.get_profile().unique_id()] << std::endl;
}
}
};
zxxx@nx:/dev$ ls video* -al
crw-rw-rw-+ 1 root plugdev 81, 0 5月 20 13:31 video0
crw-rw-rw-+ 1 root plugdev 81, 1 5月 20 13:31 video1
crw-rw-rw-+ 1 root plugdev 81, 2 5月 20 13:31 video2
2. run the callback program, the callback function is invoked, even though there are some error info
20/05 13:32:48,481 INFO [548514643808] (sensor.cpp:1340) Request: Z16 Depth, Resolved to: Z16 Depth, 20/05 13:32:48,553 INFO [548514643808] (uvc-streamer.cpp:28) endpoint 130 read buffer size: 1844224 20/05 13:32:48,596 INFO [548514643808] (sensor.cpp:1340) Request: RGB8 Color, Resolved to: YUYV Color, 20/05 13:32:48,658 INFO [548514643808] (uvc-streamer.cpp:28) endpoint 132 read buffer size: 1844224 RealSense callback sample
20/05 13:32:48,720 ERROR [548451037632] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:32:48,744 ERROR [548130419136] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:32:48,774 ERROR [548130419136] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:32:48,808 ERROR [548130419136] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:32:48,838 ERROR [548130419136] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:32:48,850 WARNING [548434252224] (messenger-libusb.cpp:42) control_transfer returned error, index: 768, error: No data available, number: 61 20/05 13:32:48,911 WARNING [548434252224] (messenger-libusb.cpp:42) control_transfer returned error, index: 768, error: No data available, number: 61 20/05 13:32:48,982 ERROR [548451037632] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:32:48,993 WARNING [548434252224] (messenger-libusb.cpp:42) control_transfer returned error, index: 768, error: Resource temporarily unavailable, number: 11 20/05 13:32:49,064 ERROR [548130419136] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:32:49,077 WARNING [548434252224] (messenger-libusb.cpp:42) control_transfer returned error, index: 768, error: Resource temporarily unavailable, number: 11 20/05 13:32:49,148 ERROR [548130419136] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:32:49,159 WARNING [548434252224] (messenger-libusb.cpp:42) control_transfer returned error, index: 768, error: Resource temporarily unavailable, number: 11 20/05 13:32:49,230 ERROR [548451037632] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:32:49,256 ERROR [548101546432] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:32:49,267 WARNING [548434252224] (messenger-libusb.cpp:42) control_transfer returned error, index: 768, error: Resource temporarily unavailable, number: 11 20/05 13:32:49,338 ERROR [548130419136] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:32:49,350 WARNING [548434252224] (messenger-libusb.cpp:42) control_transfer returned error, index: 768, error: Resource temporarily unavailable, number: 11 20/05 13:32:49,410 WARNING [548434252224] (messenger-libusb.cpp:42) control_transfer returned error, index: 768, error: Resource temporarily unavailable, number: 11 20/05 13:32:49,481 ERROR [548101546432] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:32:49,493 WARNING [548434252224] (messenger-libusb.cpp:42) control_transfer returned error, index: 768, error: Resource temporarily unavailable, number: 11 20/05 13:32:49,564 ERROR [548130419136] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:32:49,575 WARNING [548434252224] (messenger-libusb.cpp:42) control_transfer returned error, index: 768, error: Resource temporarily unavailable, number: 11 20/05 13:32:49,646 ERROR [548101546432] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:32:49,660 WARNING [548434252224] (messenger-libusb.cpp:42) control_transfer returned error, index: 768, error: Resource temporarily unavailable, number: 11 20/05 13:32:49,731 ERROR [548451037632] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:32:49,752 ERROR [548130419136] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:32:49,763 WARNING [548434252224] (messenger-libusb.cpp:42) control_transfer returned error, index: 768, error: Resource temporarily unavailable, number: 11 20/05 13:32:49,834 ERROR [548101546432] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:32:49,856 ERROR [548451037632] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:32:49,877 ERROR [548130419136] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:32:50,026 ERROR [548130419136] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:32:50,048 ERROR [548101546432] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:32:50,070 ERROR [548130419136] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:32:50,093 ERROR [548101546432] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:32:50,123 ERROR [548451037632] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:32:50,144 ERROR [548130419136] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:32:50,165 ERROR [548101546432] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:32:50,187 ERROR [548130419136] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) come into callback... counters = 1 counters = 1 20/05 13:32:50,208 ERROR [548101546432] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:32:50,229 ERROR [548130419136] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) come into callback... counters = 2 counters = 2 come into callback... counters = 3 counters = 3 come into callback... counters = 4 counters = 4 come into callback... counters = 5 counters = 5 20/05 13:32:50,251 ERROR [548451037632] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:32:50,272 ERROR [548101546432] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:32:50,295 ERROR [548130419136] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) come into callback... counters = 6 counters = 6 come into callback... counters = 7 counters = 7 come into callback... counters = 8 counters = 8 come into callback... counters = 9 counters = 9 20/05 13:32:50,316 ERROR [548101546432] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) come into callback... counters = 10 counters = 10
3. When the callback program is still running, the video devices are not in the /dev directory
zxxx@nx:/dev$ ls video -al ls: cannot access 'video': No such file or directory
4. Close the program then re-run it, callback function is never invoked.
zxxx@nx:~/workspace/local/nx_d415_test/build$ ./camera 20/05 13:46:44,385 INFO [548389916512] (sensor.cpp:1340) Request: Z16 Depth, Resolved to: Z16 Depth, 20/05 13:46:44,456 INFO [548389916512] (uvc-streamer.cpp:28) endpoint 130 read buffer size: 1844224 20/05 13:46:44,499 INFO [548389916512] (sensor.cpp:1340) Request: RGB8 Color, Resolved to: YUYV Color, 20/05 13:46:44,561 INFO [548389916512] (uvc-streamer.cpp:28) endpoint 132 read buffer size: 1844224 RealSense callback sample
20/05 13:46:44,620 ERROR [548326310336] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:46:44,742 ERROR [548326310336] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:46:44,863 ERROR [548326310336] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:46:44,984 ERROR [548326310336] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:46:45,106 ERROR [548326310336] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs) 20/05 13:46:45,228 ERROR [548326310336] (types.h:307) get_device_time_ms() took too long (more then 2 mSecs)
So, as the pipeline is still running for the 1st time, the video devices are already disappeared(seen from another terminal), which leads to the program can not be restart.
Looking forward to fixing this bug, which is important to us
@7oud Thanks so much for running the test and generating information for the developers to be able to analyze!
I can reproduce the issue with D435i. Video devices disappeared after the pipline is established for the first time.
@ancorasir As the issue is reproducible, I will label this discussion as "Bug" so that it is trackable by the RealSense developers. Thanks for your report!
Required Info | |
---|---|
Camera Model | { D400 } |
Firmware Version | (05.12.03.00) |
Operating System & Version | { Ubuntu 18.04 for Tegra |
Kernel Version (Linux Only) | (Linux nx 4.9.140-tegra) |
Platform | NVIDIA Jetson Nano, JetPack 4.4 |
SDK Version | { 2.34.1 } |
I solved this problem by downloading the latest version of v2.34.1, which runs smoothly on the Jetson Nano. @7oud @ancorasir You can try.
If you have CUDA dev-kit installed, don't forget to add -DBUILD_WITH_CUDA=true for better performance. So my cmake is:
cmake \
-DBUILD_EXAMPLES=true \
-DFORCE_LIBUVC=true \
-DBUILD_WITH_CUDA=true \
..
@shaoxiang I'm happy to hear that you had success with your Nano!
@shaoxiang @MartyG-RealSense Yes, the libRS worked well on Nano from v2.23, I have tested it. However it didn't work on Xavier NX now
@7oud Yes, I know NX has a distinct process from that of Nano to work out. We can be glad for @shaoxiang that their own problem is solved though. smiles
@shaoxiang so the problem you met on nano was fixed in v2.34.1? is it bcz of jetpack4.4 and what's the problem? when I used nano with d415, I found the conversion from yuyv to rgb with cuda was slow, conversion using cpu-based opencv was much faster
@7oud Yes, fixed by v2.34.1 and this version works well on the nano not found jam or delay at all. When the version lower than v2.34.1, I have met many problems,like: get_device_time_ms() took too long (more then 2 mSecs) and Serious frame loss! In Ros I use two cameras (d435i and d410) at the same time and get 30FPS Realtime. I will continue to test to see if there are any other problems.
I can get 435i running on Jetson NX e.g. with realsense-viewer However, after closing it will become unavailable until further reboot; Moreover, It would throw
rs-hello-realsense
RealSense error calling rs2_pipeline_wait_for_frames(pipe:0x557f97e110):
Frame didn't arrive within 15000
I wasn't able also to confirm anything but rs-distance and rs-capture; Trying to run it from ROS. @shaoxiang Are you using Jetson nano or Jetson NX? With the latter Jetpack 4.4? How do you install librealsense2 driver and examples? How do you construct ROS? ROS Dockerfile?
when the latest librealsense will be added to apt-repository?
apt policy librealsense*
librealsense2-dbg:
Installed: (none)
Candidate: 2.32.1-5ubuntu5~1.gbp83f952
Version table:
2.32.1-5ubuntu5~1.gbp83f952 500
500 http://realsense-hw-public.s3.amazonaws.com/Debian/apt-repo bionic/main arm64 Packages
2.31.0-5ubuntu5~1.gbp95d195 500
500 http://realsense-hw-public.s3.amazonaws.com/Debian/apt-repo bionic/main arm64 Packages
2.30.0-5ubuntu5~1.gbp9f99fa 500
500 http://realsense-hw-public.s3.amazonaws.com/Debian/apt-repo bionic/main arm64 Packages
2.28.1-5ubuntu5~1.gbpc1c97e 500
500 http://realsense-hw-public.s3.amazonaws.com/Debian/apt-repo bionic/main arm64 Packages
librealsense2-dev:
Installed: 2.32.1-5ubuntu5~1.gbp83f952
Candidate: 2.32.1-5ubuntu5~1.gbp83f952
Version table:
*** 2.32.1-5ubuntu5~1.gbp83f952 500
500 http://realsense-hw-public.s3.amazonaws.com/Debian/apt-repo bionic/main arm64 Packages
100 /var/lib/dpkg/status
2.31.0-5ubuntu5~1.gbp95d195 500
500 http://realsense-hw-public.s3.amazonaws.com/Debian/apt-repo bionic/main arm64 Packages
2.30.0-5ubuntu5~1.gbp9f99fa 500
500 http://realsense-hw-public.s3.amazonaws.com/Debian/apt-repo bionic/main arm64 Packages
2.28.1-5ubuntu5~1.gbpc1c97e 500
500 http://realsense-hw-public.s3.amazonaws.com/Debian/apt-repo bionic/main arm64 Packages
librealsense2:
Installed: 2.32.1-5ubuntu5~1.gbp83f952
Candidate: 2.32.1-5ubuntu5~1.gbp83f952
Version table:
*** 2.32.1-5ubuntu5~1.gbp83f952 500
500 http://realsense-hw-public.s3.amazonaws.com/Debian/apt-repo bionic/main arm64 Packages
100 /var/lib/dpkg/status
2.31.0-5ubuntu5~1.gbp95d195 500
500 http://realsense-hw-public.s3.amazonaws.com/Debian/apt-repo bionic/main arm64 Packages
2.30.0-5ubuntu5~1.gbp9f99fa 500
500 http://realsense-hw-public.s3.amazonaws.com/Debian/apt-repo bionic/main arm64 Packages
2.28.1-5ubuntu5~1.gbpc1c97e 500
500 http://realsense-hw-public.s3.amazonaws.com/Debian/apt-repo bionic/main arm64 Packages
librealsense2-gl-dbg:
Installed: (none)
Candidate: 2.32.1-5ubuntu5~1.gbp83f952
Version table:
2.32.1-5ubuntu5~1.gbp83f952 500
500 http://realsense-hw-public.s3.amazonaws.com/Debian/apt-repo bionic/main arm64 Packages
2.31.0-5ubuntu5~1.gbp95d195 500
500 http://realsense-hw-public.s3.amazonaws.com/Debian/apt-repo bionic/main arm64 Packages
2.30.0-5ubuntu5~1.gbp9f99fa 500
500 http://realsense-hw-public.s3.amazonaws.com/Debian/apt-repo bionic/main arm64 Packages
2.28.1-5ubuntu5~1.gbpc1c97e 500
500 http://realsense-hw-public.s3.amazonaws.com/Debian/apt-repo bionic/main arm64 Packages
librealsense2-gl-dev:
Installed: (none)
Candidate: 2.32.1-5ubuntu5~1.gbp83f952
Version table:
2.32.1-5ubuntu5~1.gbp83f952 500
500 http://realsense-hw-public.s3.amazonaws.com/Debian/apt-repo bionic/main arm64 Packages
2.31.0-5ubuntu5~1.gbp95d195 500
500 http://realsense-hw-public.s3.amazonaws.com/Debian/apt-repo bionic/main arm64 Packages
2.30.0-5ubuntu5~1.gbp9f99fa 500
500 http://realsense-hw-public.s3.amazonaws.com/Debian/apt-repo bionic/main arm64 Packages
2.28.1-5ubuntu5~1.gbpc1c97e 500
500 http://realsense-hw-public.s3.amazonaws.com/Debian/apt-repo bionic/main arm64 Packages
librealsense2-gl:
Installed: 2.32.1-5ubuntu5~1.gbp83f952
Candidate: 2.32.1-5ubuntu5~1.gbp83f952
Version table:
*** 2.32.1-5ubuntu5~1.gbp83f952 500
500 http://realsense-hw-public.s3.amazonaws.com/Debian/apt-repo bionic/main arm64 Packages
100 /var/lib/dpkg/status
2.31.0-5ubuntu5~1.gbp95d195 500
500 http://realsense-hw-public.s3.amazonaws.com/Debian/apt-repo bionic/main arm64 Packages
2.30.0-5ubuntu5~1.gbp9f99fa 500
500 http://realsense-hw-public.s3.amazonaws.com/Debian/apt-repo bionic/main arm64 Packages
2.28.1-5ubuntu5~1.gbpc1c97e 500
500 http://realsense-hw-public.s3.amazonaws.com/Debian/apt-repo bionic/main arm64 Packages
librealsense2-utils:
Installed: 2.32.1-5ubuntu5~1.gbp83f952
Candidate: 2.32.1-5ubuntu5~1.gbp83f952
Version table:
*** 2.32.1-5ubuntu5~1.gbp83f952 500
500 http://realsense-hw-public.s3.amazonaws.com/Debian/apt-repo bionic/main arm64 Packages
100 /var/lib/dpkg/status
2.31.0-5ubuntu5~1.gbp95d195 500
500 http://realsense-hw-public.s3.amazonaws.com/Debian/apt-repo bionic/main arm64 Packages
2.30.0-5ubuntu5~1.gbp9f99fa 500
500 http://realsense-hw-public.s3.amazonaws.com/Debian/apt-repo bionic/main arm64 Packages
2.28.1-5ubuntu5~1.gbpc1c97e 500
500 http://realsense-hw-public.s3.amazonaws.com/Debian/apt-repo bionic/main arm64 Packages
librealsense2-udev-rules:
Installed: 2.32.1-5ubuntu5~1.gbp83f952
Candidate: 2.32.1-5ubuntu5~1.gbp83f952
Version table:
*** 2.32.1-5ubuntu5~1.gbp83f952 500
500 http://realsense-hw-public.s3.amazonaws.com/Debian/apt-repo bionic/main arm64 Packages
100 /var/lib/dpkg/status
2.31.0-5ubuntu5~1.gbp95d195 500
500 http://realsense-hw-public.s3.amazonaws.com/Debian/apt-repo bionic/main arm64 Packages
2.30.0-5ubuntu5~1.gbp9f99fa 500
500 http://realsense-hw-public.s3.amazonaws.com/Debian/apt-repo bionic/main arm64 Packages
2.28.1-5ubuntu5~1.gbpc1c97e 500
500 http://realsense-hw-public.s3.amazonaws.com/Debian/apt-repo bionic/main arm64 Packages
I'm experiencing similar issues as described throughout this thread. With a D435i, a Xavier NX, and version 2.32 (installed via apt-get), realsense-viewer is able to open the depth and color streams, but individually they stop working after being closed and re-opened, and after this the video devices have disappeared from /dev.
As a sanity check, I've confirmed it's not a hardware issue. The sensor works fine on my x86_64 Ubuntu laptop (with the same library version installed from apt) and my Windows machine.
@MasonLilly If one stream type is stopped and another type is started and then that is stopped and the previous type was started, after several open-closes of different stream types (instead of doing stop-start on the same stream type), there is a chance that the SDK may become confused.
The latest 400 Series camera firmware 5.12.5 contains a fix for the problem of IR frames no longer arriving after several stop-starts.
I have this problem even if i toggle a single stream on and off. In my case, it almost always occurs after the first on/off cycle.
I somewhat doubt that it's a firmware issue, since I've had no problem running the same camera with the same firmware on Windows and non-Jetson Ubuntu, but I'll try the different firmware soon and see.
On Fri, Jun 5, 2020, 2:33 AM MartyG-RealSense notifications@github.com wrote:
@MasonLilly https://github.com/MasonLilly If one stream type is stopped and another type is started and then that is stopped and the previous type was started, after several open-closes of different stream types (instead of doing stop-start on the same stream type), there is a chance that the SDK may become confused.
The latest 400 Series camera firmware 5.12.5 contains a fix for the problem of IR frames no longer arriving after several stop-starts.
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/IntelRealSense/librealsense/issues/6412#issuecomment-639287212, or unsubscribe https://github.com/notifications/unsubscribe-auth/ADBJ3PFPVW4EEZVPNVSZVZ3RVCGSDANCNFSM4ND6EAUA .
Please let us know the results of a firmware update, @MasonLilly and if it does not help then I can look at filing an official bug report, since a few people in this discussion seem to be having a similar problem.
I'll be darned, that seems to have worked! Since updating the firmware I haven't had the above issue, no matter how capriciously I turn streams on and off. Many thanks!
Great news, I'm very pleased for you :) Is it okay to close the case now if you are satisfied with the outcome? Thanks!
Yes, the latest firmware with v2.35 SDK works normally. Thanks!
Had the same issue of No Frames Received
from the rgb camera with a D435i (5.12.3) on librealsense 2.33.1 on a Xavier NX with Jetpack 4.4.
I’ve updated the camera to 5.12.5 and it seems to work fine now. I see a white screen for a second when restarting the RGB stream and then the picture appears. Don’t know what’s causing this but don’t remember it happening with a regular nano.
Also the control transfer returned error
message still appears at the start when launching the depth stream.
@jorgemia I don't know if you are using ROS, but the opinion of Doronhi the RealSense ROS developer has been that if the control_transfer returned warning is only appearing at the start and is not generating continuously then it is not something to be concerned about. He said:
"The warning message "control_transfer returned error" reflects communication error with the device that is resolved by repeating the attempt. These messages are still under investigation but as long as they appear only now and then, say at the beginning of the run, they can safely be ignored for now".
"If the message "messenger-libusb.cpp:42) control_transfer returned error, index: 768, error: Resource temporarily unavailable, number: 11" is continuous, that is appears multiple times per second, then it is not good. It suggest a real communication problem between the code and the device and can have effects".
Yes already knew about it, thanks @MartyG-RealSense !!
Issue Description
I installed realsense SDK on the latest NVIDIA Jetson Xavier NX with D415, as installation_jetson.md, however, it didn't work. The realsense viewer prompted "No Frames Received" most of time and showed images few times. The same installation method worked on Jetson Nano.
How to deal with the new hardware(Jetson NX), building from source can fix this problem? or more complicated method should be used? Please give some advice