LVS
lvs-users
Google
 
Web LinuxVirtualServer.org

[lvs-users] keepalived+bridging!!

To: lvs-users@xxxxxxxxxxxxxxxxxxxxxx
Subject: [lvs-users] keepalived+bridging!!
From: "Ramsurrun Visham" <vishamr@xxxxxxxxx>
Date: Wed, 20 Jun 2007 21:46:57 +0400
Hi to all,

My topology is as follows:

         |                                         |
         |     VIP 192.168.0.10           |
         |                                         |
         | br1 192.168.0.2                  | br1 192.168.0.3
         | eth1                                  | eth1
     |Master| -------  eth2 --------- |Backup|
  eth0 |                                         | eth0
br0 192.168.10.2                  br0 192.168.10.3
         |                                         |
         |                                         |
         |     VIP 192.168.10.10         |
         |                                         |

The IP addresses of eth2 on Master is 192.168.100.100 and on Backup is
192.168.100.200.

My keepalived.conf on the MASTER node is as follows:
--------------------------------------
! Configuration File for keepalived

vrrp_sync_group G1 {   # must be before vrrp_instance declaration
 group {
   VI_1
 }
}

vrrp_instance VI_1 {
   state MASTER
   interface eth2
   virtual_router_id 51
   priority 100
   advert_int 2
   authentication {
     auth_type PASS
     auth_pass visham
   }
   virtual_ipaddress {
       192.168.10.10/24 dev br0
   }
}
--------------------------------------

The vrrp adverts are multicast along eth2. My problem is that I don't
know how how set the virtual ip address on br1 since I have already
defined a vrrp_instance for eth2. Is it possible to several virtual ip
addresses in one vrrp_instance?

   virtual_ipaddress {
       192.168.10.10/24 dev br0
       192.168.0.10/24 dev br1
   }
}

Unfortunately, I dont have a PC equipped with 3 NICs to test it.

Also, I wanted to know if virtual MAC (VMAC) is supported by
keepalived. If yes, how do I do it because I want the BACKUP PC to
have the same MAC address as the MASTER when the MASTER node fails.

Thx in advance...

Warm regards,
Visham


<Prev in Thread] Current Thread [Next in Thread>