Chris Ross <cross+netbsd%distal.com@localhost> writes: > On May 12, 2010, at 4:13 PM, Greg Troxel wrote: >> Try 'route -n monitor' to a file while getting both an automatic and a >> manual address and see what happens. > > The route -n monitor running from rc.local causes the later start of > ntpd to not produce error messages. But, it seems that fstat still > reports the ntpd binding to the global address as after the route > socket, so I presume that route(8) is not causing the address to hit > earlier, it's just for some reason changing the way ntpd messes up. > > Any additional thoughts? I can let you know what's in the route -n > monitor output, but i'm not sure that it's relevant at this point. That's odd, but if it's consistent that running route monitor avoids the error that itself is interesting.
Description: PGP signature