There are 3 main way for the device to discover the ztp server:
- DHCP Option 43: I hate with passion because the documentation isn't super clear and I'm using it for AP discovery, therefore I didn't tested it
- DNS: The device will look for pnpserver.localdomain, I like this one better and it is what I ended up using
- Cloud redirect: couldn't test it because my lab sits behind a proxy
This is the DHCP configuration needed, note the DNS server and the local domain.
ip dhcp pool Vlan1
network 10.102.1.0 255.255.255.0
default-router 10.102.1.1
dns-server 10.102.2.20
domain-name secmob.local
(and remember to create an entry on the DNS server!!)
- Before everest - 16.5.1a - no discovery at all is done using that interface
- in 16.5.1a the interface will be down after the pnp operation
interface GigabitEthernet0/0
vrf forwarding Mgmt-vrf
ip address 10.102.1.48 255.255.255.0
negotiation auto
shutdown
I have tried to force the no shutdown
on the configuration, but this in turn made the pnp process fail AND shutted down the port anyway - investigations in process
2- In 16.5.1a weird behaviours are observed if the configuration for the port used for pnp is not in switchport
In my case this is the intended configuration:
interface GigabitEthernet1/0/37
no switchport
ip address 10.102.1.11 255.255.255.0
And this is the outcome:
interface GigabitEthernet1/0/37
no switchport
no ip address
Workaround: use switcport and move the ip to the vlan.