SUMMARY: nic failover

Thanks to the following people for their quick and
helpful response. This is sufficient info for me to
proceed with the IP multipathing. Once again, Thanks -
this list is great!!!

Giovanni Navarrette
Doug Otto
Kevin Buterbaugh
Rasmussen Thorfinn
Darren Dunham
ChinTu
Dirk Bonning
Muralidharan P
Wanke Matthias

nic failover for Solaris 8 is also supported. Useful
reference and information were given as follow

1) IP Network Multipathing in Solaris 8 (forwarded by
Giovanni Navarrette)
http://www.eng.auburn.edu/users/doug/howtos/multipathing.html

2)Excerp newsletter forwarded by Doug Otto

Vol 3.1         TechTips - A publication of Govstor,
LLC.       03/2002

+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
                               GovStor
                    Systems, Storage and Networking
                       Solutions for Government.
                           1319 20th Street
                        Sacramento, CA  95814
                            (916)446-3232
                          sales@govstor.com
                       http://www.govstor.com/
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
                   Network Failover with Solaris 8
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

For this article I'm going to continue my theme of
introducing features
in Solaris 8 that you might not be aware of.  Included
with the 10/00
release of Solaris 8, (MU2 for those of you who've
been paying
attention), are some handy configuration options to
the ifconfig command
collectively known as IP Multipathing.

Along with the new configuration options comes a new
daemon called
in.mpathd.  The application lives in /sbin, right
where you'd expect it
to be, and starts automatically unless the
configuration file
/etc/default/mpathd has been removed.

As with most things UNIX, there are a numbers of
configuration options
available.  To avoid confusion, I'm only going to
explain a simple
failover scenario using only two NICs for the purpose
of this article.

To behave predictably, each NIC in your failover group
needs to have a
unique MAC address.  In most cases this can be
accomplished by setting
local-mac-address?=true from the OBP or via the eeprom
command; either
way a reboot is in order.

If your failover NIC does not have a local MAC address
like the older
Lance Ethernet adapters (le) you will have to give it
one in a startup
script (ifconfig le0 ether xx.xx.xx.xx.xx.xx).  MAC
addresses need to be
unique on your network so check your ARP table (arp
-a) before
arbitrarily assigning a new address.

To start configuring my system we need to make some
host file additions.
Each NIC that we want to failover requires a test
address to be
associated with it.  We also need to assign a valid IP
to the "backup"
NIC:

127.0.0.1       localhost loghost
192.168.0.4     wizard                  # Primary
Interface
192.168.0.7     wizard-test             # Test address
for IPMP
192.168.0.8     wizard-hme1             # Failover for
IPMP

The next step is to create the appropriate
/etc/hostname.* files for
each interface.  On my test system these files are
/etc/hostname.hme0
and /etc/hostname.hme1.

In the first file we assign the primary interface to a
group called
"ipmp" and bring it up with the failover option set. 
We also add hme0:1
with the addif option, tell the operating system not
to use it for
applications or failover and bring it up.

/etc/hostname.hme0:
wizard group ipmp failover up addif wizard-test
deprecated -failover up

The next task is to configure the hme1 interface to
take over in the
event of a failure on hme0.   The command adds the
interface to the
"ipmp" group, assigns a MAC address, tells the
operating system not to
use it for applications, not to fail it over when an
error is detected
and to use it as a standby.


/etc/hostname.hme1:
wizard-hme1 group ipmp deprecated -failover standby up

After a system boot, the output of ifconfig -a looks
roughly like the
following:

-----
lo0: flags=1000849<UP,LOOPBACK,RUNNING,MULTICAST,IPv4>
mtu 8232 index 1
        inet 127.0.0.1 netmask ff000000
hme0:
flags=1000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4> mtu
1500 index
2
        inet 192.168.0.4 netmask ffffff00 broadcast
192.168.0.255
        groupname ipmp

hme0:1:
flags=9040843<UP,BROADCAST,RUNNING,MULTICAST,DEPRECATED,IPv4,NOFAILOVER>
mtu 1500 index 2
        inet 192.168.0.7 netmask ffffff00 broadcast
192.168.0.255

hme1:
flags=79040843<UP,BROADCAST,RUNNING,MULTICAST,DEPRECATED,IPv4,NOFAILOVER
,STANDBY,INACTIVE> mtu 1500 index 3
        inet 192.168.0.8 netmask ffffff00 broadcast
192.168.0.255
        groupname ipmp



For a basic two NIC failover group, that's all there
is to it!  Within
10 seconds of a detected failure on hme0, hme1 will
take over.
Switch-over time can be adjusted but the smaller the
number the more
"test" traffic on your network.

As always, for more detailed information refer to your
system man pages
and the resources available to you on docs.sun.com 
and
www.sun.com/blueprints.


Original Question:

Hi List,
>
> Pardon me for this question as Im a newbie on this
> platform. I heard that nic failover is one 
> of the features of Solaris 9. But Im not sure
whether
> if this is also a feature in Solaris 8. Appreciate
if
> someone could clarify this.....
>
> Thanks in advance,
>
> Nicole
Win a DV camera, join our "30 Seconds of Fame" contest   
http://sg.movies.yahoo.com
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers