I’ve got the below route table on my windows machine, the 10.235.0.1 interface is my ZeroTier connection and I’m trying to reach various devices on the 10.235.0.0/16 network.
I can’t connect with any devices on the ZeroTier (10.235.0.0/16) network and if I do a tracert on an address, Windows always chooses my default gateway on the local network (172.16.0.19). This doesn’t seem right as there is a more specific route in the table (10.235.0.0/16 via 10.235.0.1).
Can anyone suggest how/why Windows might be refusing to use the more specific route?
I don’t know if it’s related in some way but even after re-installing ZeroTier, my system can’t find the zerotier-cli command either with or without administrator privileges.
I’ve done some research and worked out what was going on.
It seems that Windows (any maybe other OS’s?) sends out ARP requests on the correct interface according to the routing table but if they don’t come back then it reverts to the default route.
There was a mistake in my ZT flow rules so ARP wasn’t working, which was causing the reversion to the default route.
I have some old clients (v1.1.14 - I have Teltonika being years behind OpenWRT to thank for these) which don’t support tags in the flow rules.
I was depending on tags to allow ARP and a few other things so I created some separate rules to allow it and once windows could successfully ARP the device via the correct route, it used it rather than reverting back to the default route.
