dhcpcd-discuss

Hi,

I'm using dhcpcd in a virtual environment with DHCPv6 prefix delegation. 
There is a constraint of which I'm unsure as to whether this does 
influence dhcpcd's timeout behavior, so I would be glad if someone here 
can shed some light on it. I'm also unsure as to whether I simply missed 
something in the general design and configuration of dhcpcd.

The use case is as follows: I have a router cascade of two IPv6 routers, 
each with dhcpcd. On each router, eth0 faces upstream, eth1 downstream. 
Due to the environment, both routers run as docker containers, and a 
side effect of all this is that et0 and eth1 are more or less 
permanently up (if that is of any importance).

Sometimes, the lower router dhcpcd becomes ready before the upper router 
gets ready with its local DHCPv6 PD server. What I notice is that dhcpcd 
tries to contact the upper DHCPv6 server for some time and then falls 
silent. dhcpcd does not terminate, but it doesn't seem to retry 
contacting the server again -- unless I'm missing here some 
communication by not waiting long enough.

1. So, do I need statically IP configuration in order to make dhcpcd 
fall back on it and terminate after configuring the fallback?

2. Does dhcpcd also time out and terminate without a fallback 
configuration? If this is possible, how do I configure this -- 
preferably without a fallback configuration (unless this could be an 
empty one)?

My rationale here is that I would like to make the routers work without 
having to up/down interfaces if possible, and without any dependy on the 
startup sequence.

Best regards,
Harald