From owner-freebsd-hackers Fri Aug 16 09:20:50 1996 Return-Path: owner-hackers Received: (from root@localhost) by freefall.freebsd.org (8.7.5/8.7.3) id JAA26235 for hackers-outgoing; Fri, 16 Aug 1996 09:20:50 -0700 (PDT) Received: from dg-rtp.dg.com (dg-rtp.rtp.dg.com [128.222.1.2]) by freefall.freebsd.org (8.7.5/8.7.3) with SMTP id JAA26208 for ; Fri, 16 Aug 1996 09:20:43 -0700 (PDT) Received: by dg-rtp.dg.com (5.4R3.10/dg-rtp-v02) id AA20777; Fri, 16 Aug 1996 12:20:04 -0400 Received: from ponds by dg-rtp.dg.com.rtp.dg.com; Fri, 16 Aug 1996 12:20 EDT Received: from lakes.water.net (lakes [10.0.0.3]) by ponds.UUCP (8.7.5/8.7.3) with ESMTP id BAA01328; Fri, 16 Aug 1996 01:52:12 -0400 (EDT) Received: (from rivers@localhost) by lakes.water.net (8.7.5/8.6.9) id LAA03762; Fri, 16 Aug 1996 11:58:32 -0400 (EDT) Date: Fri, 16 Aug 1996 11:58:32 -0400 (EDT) From: Thomas David Rivers Message-Id: <199608161558.LAA03762@lakes.water.net> To: ponds!mt.sri.com!nate Subject: Re: NFS and the nameserver... Cc: ponds!freefall.cdrom.com!freebsd-hackers Content-Type: text Sender: owner-hackers@FreeBSD.ORG X-Loop: FreeBSD.org Precedence: bulk > > We need a small change to /etc/rc. > > And I assume you've made that change. :) > > > I discovered this when I set up a primary nameserver on a machine > > that also wanted to mount NFS mount points. > > > > If you have an /etc/resolv.conf, and you are the primary name > > server. The > > > > mount -a -t nfs > > > > in /etc/rc happens _before_ named is started. This means that > > the mounts fail (timing out eventually). > > > > I think it would be prudent to move the > > > > mount -a -t -nfs > > > > command until after all the networking deamons have begun, but > > particularly after named has started. > > Except that some of the networking daemons live in /usr, which must > mounted on many boxes which boot diskless and/or mount /usr alone. > > There is no 'perfect' solution to where the NFS mount commands go, and > it was determined that the current position is the 'best' for most > people. If it doesn't work for your installation, move it (which you've > done already). > > > > Nate > Well - yes, err..., umm... I made that change. Seems like a simple 'if' is appropriate here. You could set up a sysconfig parm, i.e. "diskless" and do the NFS mounts early if this is a diskless machine. if [ ! X"$diskless" = X"NO" ] then mount -a -t nfs NFS_mounted=yes fi .... after named, etc... if [ ! "NFS_mounted" = "yes" ] then mount -a -t nfs fi also - another clever trick is to put the mount command early in /etc/rc and simply run it in the background. Then, it's likely that the nameserver will get "cranked up" before the timeout and the mounts will succeed... - Just a thought - - Dave Rivers -