Thanks to:
Harvey Wamboldt
mvanderv@yahoo.com
Ronald Loftin
Sam Vilain
Jonathan Loh
Arvinder Paul
marka@bluestone.com
Tim McDonough
Anthony Worrall
Benjamin Cline
Seth Rothenberg
MARC.NEWMAN@chase.com
Tom Ballingall
Joselito Alfonso
Jim Hebert
My original question:
> Just a quick question. Does Solaris support backup Ethernet interface
> (i.e., having a backup interface take over with the same IP address if
> the primary interface stops working)?
Several people suggest Sun's "Alternate Pathing". It could be found at
"http://docs.sun.com". (search by the keyword "Alternate Pathing").
However, it won't do automatic failover. You have to switch it over manually.
Arvinder Paul pointed out that a product called Etherchannel by Sun could
do this (I haven't checked this yet); Tom Ballingall said that they are
using "Sun Trunking 1.0.1" with a quad ether card on their Enterprise 450's.
Sam Vilain and Seth Rothenberg think it's easy to be done by a simple
script. The following is the script sent by Seth Rothenberg (thanks you
again):
==========================================================
#!/bin/sh
# script to add static route to configure le1
if [ ! -d /usr/bin ]
then # /usr not mounted
exit
fi
echo "IN S71configle0" >> /tmp/success
date > /tmp/success
ifconfig -a >> /tmp/success 2>&1
echo "setting address for le1" >> /tmp/success
ifconfig le1 plumb >> /tmp/success 2>&1
ifconfig le1 ether aa:0:04:0:2:50 2>&1
ifconfig le1 inet testdg1 >> /tmp/success 2>&1
echo "setting netmask for le1" >> /tmp/success
ifconfig le1 netmask class_c broadcast + up >> /tmp/success 2>&1
ifconfig -a >> /tmp/success 2>&1
ping 255.255.255.255
netstat -r >> /tmp/success
==========================================================
The concept here is:
when the primary interface(le0, for example) stop working, then
ifconfig le0 down
ifconfig le0 unplumb
ifconfig le1 plumb
ifconfig le1 `uname -n` up
He also mention that in this case, make sure you do NOT have 2 interfaces
online with the same ethernet address.
The problem here is how to detect an interface error. I'm not a
programmer, so I don't know if there's any function which can check
whether a network interface is working properly or not(although I
think there should be one/some). To me, it seems like the command
"netstat -I le0" would do the trick. This command shows the state of the
interface. I think this might not be the best way. So please don't
hesitate to email me if you have a better resolution. I'll summarize
again.
Here are some other replies I got:
===================================
>From Harvey Wamboldt:
Yes. I'm not an expert, but one way to do this would be to install
two interfaces, then set the metric on the "backup" interface higher
than the "primary" interface. See ifconfig(1M). You should also be
aware that the ethernet address for a SPARCstation is stored in the
mainboard EEPROM. This means your startup scripts must "manually" set
the ethernet address for additional interfaces using something like:
/sbin/ifconfig hme0 ether xx:xx:xx:xx:xx:xx
(I put mine in /etc/rcS.d/S30rootusr.sh inside a test for
/etc/hostname.hme0).
===================================
>From marka@bluestone.com:
This is a feature that is currently only supported as part of the
SunCluster software; the package is called NAFO (Network Adapter
Fail-Over). Maybe you could start a movement to pressure Sun into
unbundling NAFO from the SunCluster stuff (it really does work standalone,
they just don't sell or support it that way).
===================================
>From Joselito Alfonso:
YES, Solaris does support it...we have that setup and we're using Sun
Cluster 2.1.
We have a high availability configuration that has a backup for Ethernet
interfaces
(hme0 <--> hme3 and hme1 <--> hme2 ) and of course for the array and the
system itself...
===================================
Jerry Lu
Unix Administrator
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:13:16 CDT