May-15th-2019, 12:08 PM
Thanks for the clarification. Is it in the realm of possibility that you'll consider implementing a work-around with udev, script and yaml file containing udevadm info output on the host hilscher system?
Is it possible to mount bus devices under /dev/bus/usb/001/... directly with --device? The main problem is, that devices might get reconnected and get assigned different numbers under /dev/ttyUSB{0,1,...}. Even if they are connected on the same port physically but in different order. So the idea would be to just mount the physical location, assuming the logical usb device if exchanged is always connected to the same physical port on the bus. I'd then implement a check, if the device becomes unavailable to kill the container and the restart=always rule will then restart the container until the device becomes available again. If the application is designed as a microservice architecture with a second service monitoring the status of the first one, I'll still be able to react on situations, when a device is unreachable for some reason.
Is it possible to mount bus devices under /dev/bus/usb/001/... directly with --device? The main problem is, that devices might get reconnected and get assigned different numbers under /dev/ttyUSB{0,1,...}. Even if they are connected on the same port physically but in different order. So the idea would be to just mount the physical location, assuming the logical usb device if exchanged is always connected to the same physical port on the bus. I'd then implement a check, if the device becomes unavailable to kill the container and the restart=always rule will then restart the container until the device becomes available again. If the application is designed as a microservice architecture with a second service monitoring the status of the first one, I'll still be able to react on situations, when a device is unreachable for some reason.