OH2.1 and RFXCOM takes ages to come online

I’ve just upgraded from a stable OH V2.0 into a problematic OH V2.1. My main problem is that it takes several variable minutes for my RFXCOM device to come online meaning that it spits errors when persistance restores the values to an item that doesn’t have a working bridge, which goes on to cause problems. An error is below

[WARN ] [ome.core.thing.internal.ThingManager] - Cannot delegate command 'OFF' for item 'TopStairLightsDisable' to handler for channel 'rfxcom:lighting1:10351c54:command', because no handler is assigned. Maybe the binding is not installed or not propertly initialized.

The bridge came online 7 minutes and 10 seconds after it finished loading default.items, the next time it took.5 minutes and 13 seconds, and the persistance works at this lower level of delay

The USB connection to the bridge is via a hard link, and apart from the onboard bluetooth its the only thing plugged into the Pi3

Is there a way of either hurrying this up, or delaying persistence until it announces its online status?

I’ve already cleared out the tmp and cache folders chasing a previous fault, as well as uninstalled and reinstalled the RFXCom binding

Did this only happen just after the upgrade or every time after that as well?

It’s pretty slow to start in general, the only sensible way I’ve found to make it reliable is to take a bunch of my rule files and copy them away from OpenHAB via a cronjob at system start, and then copy them back once a known item is no longer in the NULL state. It seems the RFXCOM items are consistently around the last things to come online and sending anything to them before then puts errors in the log.
It’s a pity the RFXCOM bridge being online can’t be mapped to an item that can be checked after a reboot