21:33:46 <barfield> sdc-napi documentation regarding static routes is ambiguous regarding link local syntax. /networks/<UUID>?action=update -X PUT '{ "routes": "10.0.0.0/9": "192.168.128.1", "linklocal": true }' always fails. Any advice to update sdc:routes from napi to change "linklocal": false to "linklocal": true?
21:34:51 <barfield> I'm having some issues with rh-style bhyve's setting static routes on both the external and fabric nics. 
21:35:29 <barfield> The downfall of this behavior is that the iproute2 facility applies the same source IP address of net1 to both net0 and net1, causing the routes to fail
21:36:00 <barfield> The failure is also true for the /fabrics endpoign. 
21:36:41 <barfield> They do work if I manually delete them from net0 like `ip route delete 10.0.0.0/8 via 192.168.128.1 dev net0`
21:37:05 <barfield> But that has to be done within the guest directly. It isn't a big deal with ansible but they reappear after reboot. 
21:38:15 <barfield> After re-reading the question I would like to specify, that I am attempting to have the metdata agent on the VM globalZone update the sdc:routes key from napi. 
21:40:12 <barfield> Another question that I have while I am here: how can apply "allow_ip_spoofing": true on a fabric network so that all VM's can spoof IP's if they would like on fabrics? I would never do this on a public network obviously. But I see no need to regulate IP addresses on a customers dedicated fabric. This quesiton arises because of kubernetes CNI's that route pod traffic using their own subnets. 
21:42:08 <barfield> I am sure other security concerns arise even on fabrics with allowing spoofing. But for the time being until I find a better way to allow a CNI to natively manage pod traffic this seems like the quickest solution.