Red Hat Bugzilla – Bug 49621
RFE: separate ipv6 support
Last modified: 2007-04-18 12:34:59 EDT
IPv6 support has been requested in #41309.
Unfortunately xinetd's ipv6 support hasn't been designed all that well (it fails on ipv4-only system,
among other things), and patching it would take a _lot_ of work.
Therefore I propose to add xinetd-ipv6 subpackage with only /usr/sbin/xinetd-ipv6 binary inside,
same as /usr/sbin/xinetd except with ipv6 support.
In /etc/init.d/xinetd, if you have IPv6 enabled in configuration and the subpackage installed, then
IPv6-compatible xinetd would be run.
This has been implemented in .src.rpm against 2.3.0-2 in http://www.netcore.fi/pekkas/linux/ipv6/.
Added in xinetd-2.3.0-3
One note: if you have traditional xinetd running, then enable IPv6 and do '/sbin/service xinetd restart', the
script gets confused because it can't stop 'xinetd-ipv6'. Going around this would be rather complicated
and might uglify the code, so I didn't go into it. You might want to consider whether this is an acceptable
This can be fixed by a dirty trick (also fixes status) -- reopening just in
--- xinetd.init.orig Tue Jul 24 10:10:55 2001
+++ xinetd.init Tue Jul 24 10:11:21 2001
@@ -67,7 +67,7 @@
echo -n $"Stopping $prog: "
- killproc $prog
+ killproc xinetd
rm -f /var/lock/subsys/xinetd
@@ -77,7 +77,7 @@
echo -n $"Reloading configuration: "
- killproc $prog -USR2
+ killproc xinetd -USR2
@@ -103,7 +103,7 @@
- status xinetd
+ status $prog
Fixed in xinetd-2.3.0-4 and newer