dhcpcd-discuss

Hi David

On Monday 30 May 2016 22:04:09 David Hauck wrote:
> I have a situation where an interface has an active lease. At some point
> carrier is lost on the interface and the interface goes through the
> following sequence (this is with v6.7.1):
> 
> 20160530 14:26:38.346 info daemon(dhcpcd):eth1: carrier lost
> 20160530 14:26:38.471 info daemon(dhcpcd):eth1: 60-custom: received
> reason=NOCARRIER, ip=, sn=, gw=, dns=, options=, hn= 20160530 14:26:38.471
> info daemon(dhcpcd):eth1: deleting route to 205.159.216.128/25 20160530
> 14:26:38.471 info daemon(dhcpcd):eth1: deleting default route via
> 205.159.216.130 20160530 14:26:38.479 info daemon(dhcpcd):eth1: 60-custom:
> received reason=EXPIRE, ip=, sn=, gw=, dns=, options=, hn=
> 
> Eventually, carrier is re-established:
> 
> 20160530 14:27:08.118 info daemon(dhcpcd):eth1: carrier acquired
> 20160530 14:27:08.123 info daemon(dhcpcd):eth1: 60-custom: received
> reason=CARRIER, ip=, sn=, gw=, dns=, options=, hn= 20160530 14:27:08.124
> info daemon(dhcpcd):eth1: IAID e0:53:f3:9f
> 20160530 14:27:08.794 info daemon(dhcpcd):eth1: rebinding lease of
> 205.159.216.183 20160530 14:27:13.812 info daemon(dhcpcd):eth1: 60-custom:
> received reason=EXPIRE, ip=, sn=, gw=, dns=, options=, hn= 20160530
> 14:27:13.812 info daemon(dhcpcd):eth1: soliciting a DHCP lease
> 
> However, at this point no server is reachable. My question is that at this
> point I would expect the 'timeout' value to engage; i.e., dhcpcd exit if
> noipv4ll is specified. Do I have this wrong? If so, then the startup state
> (where timeout expiring results in dhcpcd exiting) and this "timeout after
> lease lost/rebind" state result in different run states.

You need to enable the --lastlease option for this to work.
I recently added the --lastleaseextend option recently as well (only in trunk) 
which extends the lease after it expires if no DHCP server can be found,

Roy