Re: named startup & NetworkManager

[Dan Williams <dcbw redhat com>]

On Sat, 2012-08-11 at 11:06 -0400, Gene Czarcinski wrote:
> On 08/09/2012 02:25 PM, Gene Czarcinski wrote:
> > https://bugzilla.redhat.com/show_bug.cgi?id=837173
> >
> > The above bug report concerns NetworkManager NIC device initialization 
> > and named.  Yes, the bug is against bind but you should take a look at 
> > it since it directly involves NetworkManager.
> >
> > As I say in my comment, I have been able to reliably reproduce the 
> > problem as a virtual guest but not on real hardware (and I tried).
> >
> > The problem is that NetworkManager, named, and dhcpd are all started 
> > at roughly the same time (think at least within a couple seconds). 
> > dhcpd can handle the fact that when it iinitializes that there are no 
> > NICs available and when they do become available, is uses them.
> >
> > But, with respect to named, the fact that the NICs are not available 
> > makes it get everything pretty much wrong (lots and lots of error 
> > messages in the system log).  And when the devices have been 
> > initialized by NetworkManager, named does not look again. The only way 
> > to handle it is to restart named after I can login.
> >
> > I am not familiar enough with how the new initialization (systemd) 
> > works so I do not know if there is an "easy" answer.
> I have submitted a report in the Fedora bugzilla --
> 
> https://bugzilla.redhat.com/show_bug.cgi?id=847452

If there are dependencies between things here, there are two approaches:

1) make the service that depends on networking more aware of it's real
dependencies.  That means, if named is configured to listen on some
specific interface, or some specific IP address, then the  named service
should really be waiting for that address or interface to come up, not
just "networking" to start.  But that entails modifying services and
their startup scripts.

2) enable the "NetworkManager-wait-online.service" systemd service,
which will block startup until the earlier of (1) at least one network
interface has started, and (2) 30 seconds have passed.

(C&P-ed into the bug too...)

Dan