Interesting problem taxing me, and appreciate community input.
We are utilizing the above Topaz Signature pad model, in a Horizon VDI, utilizing PCoIP and Linked clones, to mostly thin client devices.
The Topaz device in question has an integrated "USB to Serial" converter, where it builds its own COM9 port for all communication with the device.
Seemingly randomly, these devices will lose connection with the View session. You can see that the hardware (indicates) it has physically lost contact with the vmware horizon server, and then if unplugged, and replugged, see when that communication resumes. It doesn't always resume cleanly though, and often, the VDI session has to be cycled to work with the device properly, again. You'll see it in the log, re-enumerate from 0x70 to 0x80.
Doing some reading, I've eliminated a suspicion that it might be pulling too much power, as it is USB powered, and with settings, have metered it down to only using .19 AMPS steady - which is nothing, really.
I have also seen where horizon uses port 32111 for the USB passthrough on horizon, and that a QOS rule in place can help with hiccups. We aren't seeing the issue with ANY other USB device, just these Topaz USB to COM signature pad devices.
I have suggested to this customer, implementing the 32111 QOS rule, and they don't want to prioritize the traffic.
Appreciate any ideas or guidance. I don't have much visibility of the horizon environment, and can only offer suggestion as it's not one we manage. The problem is sparatic at best, with the only cataylst seemingly being that it occurs more frequently when a VDI has been left for a lunch break - and even then, not with consistency.
Thanks for any help you can offer.