Recording multiple cameras using multiprocessing

[elhananby]

Hello everyone,
I am wondering - is it possible to use Python's multiprocessing Pool method to record from several different cameras on different processes? I tried to do it, but every time the code gets stuck (with no error) on the InstantCamera initialization line.
Attached is sample code (please note the cameras are on a hardware trigger):

from pypylon import pylon
import multiprocessing as mp

def record_video(device_fn):
    cam = pylon.InstantCamera(pylon.TlFactory.GetInstance().CreateDevice(device_fn))
    cam.StartGrabbing(pylon.GrabStrategy_OneByOne)
    for i in range(2500):
        grab_result = cam.RetrieveResult(5000, pylon.TimeoutHandling_ThrowException)
        if grab_result.GrabSucceeded():
            img = grabResult.Array
            cv2.imwrite("/home/elhanan_la/Videos/cam{}_frame{}.tiff".format(grab_result.GetCameraContext(), i), img)
        grab_result.Release()

def main():
    tl_factory = pylon.TlFactory.GetInstance()
    devices = tl_factory.EnumerateDevices()
    with mp.Pool(3) as p:
        p.map(record_video, [devices[i].GetFullName() for i in range(3)])
        p.close()
        p.join()

if __name__ == "__main__":
    main()

I've also tried to first initialize the cameras via the InstantCameraArray method, and then send each camera object to the pool, but then I'm getting the error:
LogicalErrorException: Node not existing (file 'genicam_wrap.cpp', line 16600)

Thanks

[salvq]

@elhananby have you had a success to record from different cameras ?