On Fri, 2008-05-30 at 14:24 -0400, Dan Williams wrote:
On Fri, 2008-05-30 at 13:59 -0400, Bill Nottingham wrote:
> Dan Williams (dcbw(a)redhat.com) said:
> > > 'After netfs' should be good enough. Although netfs stop should
possibly
> > > do lazy umounts.
> >
> > Ok, just need to bump NM a few bits later it looks like; might as well
> > be K84 to be right after messagebus.
>
> Well, NM will exit anyway when the messagebus exits, will it not?
No, because the Debian guys got all cranky and want NM to handle dbus
dropouts, because if they need to restart dbus because of a security
issue they don't want that to take down the entire machine including
networking. So there's code to poll the bus every so often and try to
reconnect if NM gets disconnected (either due to an NM error like
sending a non-UTF8 string over the bus or if dbus quits/dies).
I am very glad the debian guys got "all cranky" on this issue,
Temrinating network connections just because a component need to
resytart is plain silly.
Simo.
--
Simo Sorce * Red Hat, Inc * New York