We need to be able to control if some daemons/init scripts are shut down

[Eric Thibodeau <kyron@neuralbs.com>]

Specific conversation was over the following on IRC:

15:29 < kyron> Uber: latest works! Only stuff left is:  http://wiki.neuralbs.com/~kyron/soc2008/screenshoots/ShutdownCruft11.jpg
15:38 <@Uber> kyron: what do you exepct me to do about that?
15:43 < kyron> errors cannot exist! :P
15:44 <@Uber> I can add some knobs for that
15:44 <@Uber> rc_net_lo_shutdown="NO"
15:45 <@Uber> rc_netmount_shutdown="NO"
15:45 <@Uber> best I can do
15:45 <@Uber> or something similar
15:45 <@Uber> kyron: file a bug requesting that at bugs.marples.name so I don't forget
15:58 < kyron> Uber: np

More generally speaking, there are some services/mounts that shouldn't be shut down on ro mounted ROOTs (typically CDs and NFSroots). The sequence boils down to:

unmount remote rw shares
--> just an option for rc_netmount_shutdown="NO" isn't flexible enough since we have /home (rw) which needs to be unmounted cleanly but / (ro) which can (has to) be ignored by the rc_netmount_shutdown process
shut down services that could communicate such a state change to a server (ie: some monitoring daemons that advise the node is going down)
call halt/poweroff/reboot directly without all the other cleanup cruft

[roy]

 http://git.overlays.gentoo.org/gitweb/?p=proj/openrc.git;a=commitdiff;h=8106bea4dc542314a0b982c7169debd32a96ad36

Fixed :)