On Sun, Feb 15, 2009 at 04:35:31PM +0100, Alexey Shuvaev wrote: > On Sun, Feb 15, 2009 at 12:28:20AM -0800, Garrett Cooper wrote: > > I just updated my world to a recent snapshot (a build from last week) > > and I'm noting some parallelism / backgrounding which is really > > causing issues with my NIC and NFS mounts. I had to hit CTRL-D 5 times > > in order to get the system to come up because it couldn't resolve my > > NFS server's hostname, because the NIC wasn't up and going yet (as it > > uses the DHCP client in background mode due to the new default). > > > > Now I realize that this all ties back into the issue with the NIC > > (which I've approached Pyun about, and which I appreciate his help is > > solving issues with this buggy chipset), but is there really a need > > for parallelism at startup rc.d it can't properly detect dependencies > > with some cases like NFS mounts? > > > Me too. > I have ntpd failing to resolve dns names. > I have noticed this since appr. 1.350 of etc/defaults/rc.conf (12 days ago). > I was hoping this will go away... > > Commit log: > SVN rev 188010 on 2009-02-02 15:38:24Z by mtm > > Since, rc.d/defaultroute has the ability to wait for a > default route to show up we can turn this knob back on > without screwing subsequent daemons that expect to be > able to talk to the outside world. > > ^^^^^^^^^^^^^^^^^^^^^^^^^^^ > Seems it is not the case... > Interesting: setting background_dhclient="NO" does not > solve the issue. Maybe something else was changed? > My current 'workaroud' is synchronous_dhclient="YES" The background flag as actually added locally to maintain compatibility with the ISC version. I'd argue it's pointless in the async world and should be ripped out of the client entirely rather than being enabled by default. -- Brooks
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:39:42 UTC