Boot delay when RED has no connection

There is a long boot delay when the RED interface has no connection to the Internet ( e.g. the line is down ) or the DNS server are not reachable.

I could trace down the delay to /etc/init.d/unbound, which checks if the forwarder is reachable.

This test has the usual 20 seconds delay and with two DNS server, you get 40 seconds.

But the real delay is can_resolve_root() which tests every root DNS and gets the delay with every of them.

A solution would be to trust the admin and use the DNS forwarder without any testing in case they are not reachable.

This would also fix the problem that at present unbound is set to permissive mode when the DNS forwarders are not reachable.

Another delay is that /usr/local/bin/setaliases is called after all the scripts /etc/init.d/networking/red.up.

Adding the aliases before that could bring IPSec and the firewall up much faster.

Yes, we are aware of this and working on a solution.

This should no longer be the case with Core Update 137.

Actually I did all testing with core 138

You might have link on RED then, but still no connection.

Yes, link is up, but connection to the Internet is down.
This can be simulated by defining an invaild default route on RED.