On December 29, Kevin O'Gorman enlightened our ignorance thusly: > On Sat, Dec 29, 2001 at 08:23:30PM -0500, [EMAIL PROTECTED] wrote: > > On December 29, Kevin O'Gorman enlightened our ignorance thusly: > > > This is very odd. > > > > > > My 'named' server won't start from SysV. > > > > > > It won't start from a root shell if I try > > > # cd /etc/rc.d/rc3.d > > > # ./S10named start > > > > > > However, it will start if I do it by sourcing the file: > > > # . ./S10named start > > > > You know, of course, that S10named is just a symlink to > > /etc/rc.d/init.d/named? What happens if you invoke this script > > directly? You can also edit it, add "set -x" at the top for debugging > > output, and then evaluate the differences between the two invocations > > to see what's going wrong? > > > > Kurt > > I did that, and it's no where near as helpful as one would hope. For > one thing, the differences first show up during the sourcing of > /etc/rc.d/init.d/functions, which is (1) common to pretty much all > of the SysV files, and (b) not causing problems for any of the others.
Nothing terribly useful in here. What's in /var/log/messages? I'm sure that somehow you've fubared the configuration. I seem to recall an issue with OL 2.4 that the SysV script did not properly handle /var/run/named.pid, which led to all sorts of interesting problems starting and stopping BIND, particularly if you mixed usage of the SysV scripts and direct invocation of the BIND. Kurt _______________________________________________ Linux-users mailing list Archives, Digests, etc at http://linux.nf/mailman/listinfo/linux-users
