recover_duts: increase polling frequency

This is an architectural change: the "hook" will check
it's state and run exactly once if it's allowed.
Thus we need to check more often.

90 seconds is about 2x longer than it should ever take to have a
working network (wired ethernet).  My experience is a USB NIC will
be up and working within 20 seconds of power on. But on a bad day it
might take 15-20 seconds to load the NIC driver module and then dhcpcd
could timeout after 30 seconds. I don't know if shill is retrying
anything and I don't expect that is matters since dhcpcd retries.

TEST=Copy recovery_duts to DUT and reboot.
     Manually verify hook is invoked every minute after
     LONG_REBOOT_DELAY has expired.

Change-Id: I3da58c2378269ff1e10fb7e024dd2d4ec559eeb3
Commit-Ready: Grant Grundler <>
Tested-by: Grant Grundler <>
Reviewed-by: Richard Barnette <>
Reviewed-by: Grant Grundler <>
1 file changed