>Number:         12065
>Category:       bin
>Synopsis:       /etc/resolv.conf corrupted during DHCP lease renewal
>Confidential:   no
>Severity:       serious
>Priority:       high
>Responsible:    bin-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Sat Jan 27 21:58:00 PST 2001
>Closed-Date:
>Last-Modified:
>Originator:     Simon J. Gerraty
>Release:        1.5
>Organization:
Zen Programming...
>Environment:

System: NetBSD too 1.5L NetBSD 1.5L (TOO) #0: Thu Dec 7 01:09:02 PST 2000 root@too:/l0/NetBSD/obj/sys/compile/TOO sparc


>Description:

I think this is a /bin/sh bug, hence the priority.  I've only seen this
on sparcs running 1.5 and later, but others report the problem on 
earlier releases and other platforms (eg 1.4/i386).

dhclient script contains:

make_resolv_conf() {
echo search $new_domain_name >/etc/resolv.conf
for nameserver in $new_domain_name_servers; do
echo nameserver $nameserver >>/etc/resolv.conf
done
}

When dhclient-exit-hooks containing:

if [ "$new_ntp_servers" ]; then
case " $new_ntp_servers " in
*" $new_ip_address "*)	;;# don't frob ntp.conf on one of the servers!
*)
	for ntp in $new_ntp_servers
	do
		echo server $ntp
	done > /etc/ntp.conf
	if [ "$old_ntp_servers" != "$new_ntp_servers" ]; then
		[ -s /etc/rc_d/ntp ] && /etc/rc_d/ntp stop
		ntpdate -b -v $new_ntp_servers
		[ -s /etc/rc_d/ntp ] && /etc/rc_d/ntp start
	fi
	;;
esac
fi

is sourced by dhclient-script, the stdout from the ntpdate and 
ntpd startup scripts ends up in /etc/resolv.conf
With the above script this only happens when the ntp server list
is updated.

>How-To-Repeat:

Put something in /etc/dhclient-exit-hooks that will output to stdout.
>Fix:

An effective work around is to put
#!/bin/ksh 
at the start of dhclient-script.  This lends weight to the notion that 
this is a bug in /bin/sh

>Release-Note:
>Audit-Trail:
>Unformatted: