Hello, I encountered the following Error while running the OAK camera after a long period:

F: [global] [ 462796] [Scheduler00Thr] dispatcherResponseServe:928 no request for this response: XLINK_READ_REL_RESP 1

Is there any way to catch this exception (I use Python API) ? Any ideas would be greatly appreciated. Thanks!

Hi @mzbt
There is no way to catch this other than to put the whole execution in a try..catch block.
We currently don't expose the specific errors.

Thanks,
Jaka

3 months later

I have seen that today as well (with a rather recent firmware blob)

[coneslayer_node-2] [2024-09-28 05:05:55.299] [depthai] [debug] Initialize - finished
[coneslayer_node-2] [2024-09-28 05:05:55.462] [depthai] [debug] Resources - Archive 'depthai-bootloader-fwp-0.0.28.tar.xz' open: 5ms, archive read: 157ms
[coneslayer_node-2] [2024-09-28 05:05:56.029] [depthai] [debug] Resources - Archive 'depthai-device-fwp-9ed7c9ae4c232ff93a3500a585a6b1c00650e22c.tar.xz' open: 2ms, archive read: 736ms
[component_container-3] [INFO] [1727499957.062681135] [oak_container]: Load Library: /homeubuntu/ros2/ws_livox/install/ros_coneslayer/lib/libros_coneslayer.so
[coneslayer_node-2] [2024-09-28 05:05:57.080] [host] [debug] Device - OpenVINO version: 2022.1

…

[coneslayer_node-2] [19443010E13A7A2700] [1.1] [287.268] [system] [info] Cpu Usage - LeonOS 51.97%, LeonRT: 45.45%

average rate: 24.962
       min: 0.020s max: 0.065s std dev: 0.00919s window: 4217
[coneslayer_node-2] F: [global] [    246033] [Scheduler00Thr] dispatcherResponseServe:928       no request for this response: XLINK_READ_REL_RESP 1

So, this time around, it was only running for 5 minutes:-( But that's one of the shorter runs:-( After that, my ROS node does not produce any topic messages any longer:-(

    MRE? Sorry;-) I'm using coneslayer_openvino_2022.1_6shave.blob & coneslayer.json from mkrupczak3/Coneslayer inside a ROS2/Humble node on a RasPi5 with Ubuntu 22.04;-)

    OK, I pushed my sources to mw46d/ros_coneslayer , I hope, that helps a bit.

    Today it lasted for almost 13 minutes;-) on a stationary rover, so the cables don't move or anything;-) I'll probably have to create a 'node restarter' for now.

    Thanks for looking into it;-)

    -- Marco

    5 months later

    I have discovered the same issue on my setup. I am using three cameras on a Raspberry Pi 5, all connected via a Y-adapter and definitely receiving enough power. The application runs smoothly for about 10-15 minutes, but then one of the cameras stops working. It is always the same camera. After a restart, everything works fine again.

    If you need more detailed information about anything, I’d be happy to provide it

    It might be a thermal issue!

    The following is my last reply to a mailing list thread on that issue (from late September 2024)

    With the help of the DepthAI debug output, I found that regularly shortly after the reported chip temperatures got up to 77C, the pipeline would 'shut down' and the host would not see any data any longer.

    So, I strapped a 50mm fan to the Oak-D and it was running fine for an hour;-) The reported  chip temps stayed around 62C.

    Next step was to replace the mount and integrate two 30mm fans. Now the temps stay around 55C and the pipeline is still going strong after 1.5h 😉

    Attache are two pictures of the mount, one with the rigged up 50mm fan and the current version with the two 30mm fans;-)

    [coneslayer_node-2] [19443010E13A7A2700] [1.1] [4915.611] [system] [info] Temperatures - Average: 54.48C, CSS: 55.51C, MSS 54.42C, UPA: 55.07C, DSS: 52.90C
    [coneslayer_node-2] [19443010E13A7A2700] [1.1] [4915.611] [system] [info] Cpu Usage - LeonOS 60.79%, LeonRT: 50.47%
    [coneslayer_node-2] [19443010E13A7A2700] [1.1] [4916.613] [system] [info] Memory Usage - DDR: 66.99 / 333.28 MiB, CMX: 2.47 / 2.50 MiB, LeonOS Heap: 47.34 / 81.76 MiB, LeonRT Heap: 5.70 / 39.90 MiB / NOC ddr: 3297 MB/s

    Maybe that's helpful to some;-