How to use /etc/rc.conf with ifconfig_lagg

From: Daniel Eischen <deischen_at_freebsd.org>
Date: Mon, 21 Apr 2008 23:55:00 -0400 (EDT)
I'm trying to get a lagg interface up via standard /etc/rc.conf
settings, but it just doesn't seem to be working.  Here's the
respective settings:

   ifconfig_sis0="up"
   ifconfig_sis1="up"
   ifconfig_sis2="up"
   cloned_interfaces="lagg0"
   ifconfig_lagg0="inet 192.168.3.97 netmask 0xffffff80 \
      laggproto failover laggport sis2 laggport sis0 laggport sis1"

I've also tried:

   ifconfig_lagg0="laggproto failover laggport sis2 laggport sis0 \
      laggport sis1 inet 192.168.3.97 netmask 0xffffff80"

and without 'inet' as lagg(4) suggests:

   ifconfig_lagg0="laggproto failover laggport sis2 laggport sis0 \
      laggport sis1 192.168.3.97 netmask 0xffffff80"

but nothing seems to work.  The lagg0 interface gets created
and configured with failover and with all the failover ports
correctly added, but the IP address and netmask do not get
set.  I have to manually set them after the system boots.
Without manually setting them, I get:

lagg0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500
         options=8<VLAN_MTU>
         ether 00:06:d5:11:40:92
         media: Ethernet autoselect
         status: active
         laggproto failover
         laggport: sis1 flags=0<>
         laggport: sis0 flags=0<>
         laggport: sis2 flags=5<MASTER,ACTIVE>

After manually doing 'ifconfig inet 192.168.3.97 netmask 0xffffff80',
it works:

lagg0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500
         options=8<VLAN_MTU>
         ether 00:06:d5:11:40:92
         inet 192.168.3.97 netmask 0xffffff80 broadcast 192.168.3.127
         media: Ethernet autoselect
         status: active
         laggproto failover
         laggport: sis1 flags=0<>
         laggport: sis0 flags=0<>
         laggport: sis2 flags=5<MASTER,ACTIVE>

What's the magic to get these automatically configured and
working via /etc/rc.conf?

The other problem I have is that failover to any port
other than sis0, causes a 30+ second delay before the port
is actually usable.  For example, with cables connected
to sis0 and sis2, if I disconnect sis2, the pings continue
without interruption (failover to sis0).  If I reconnect
sis2, the pings stop for 30+ seconds before continuing.
Disconnecting sis2 again will not cause any interruption
when failing back over to sis0.  If I disconnect both
sis0 and sis2, then reconnect sis2, there is still a
30+ second delay before the pings continue.  All sis ports
do have the "short cable fix" problem.

This is under a fresh RELENG_7.  If I can I will try
HEAD as well as another system with non-sis ethernet.

-- 
DE
Received on Tue Apr 22 2008 - 01:55:02 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:39:30 UTC