|
Brad Dameron a écrit :
On Wed, 2006-06-28 at 10:22 +0100, Dr A V Le Blanc wrote:
For many years we have used keepalived to manage our lvs farm,
but I have never managed to get it to failover automatically,
so I've simply used manual failover. I have tried a large number
of minor variations on configuration files, but let me give a
simple example. There are two directors. One has this in its
/etc/keepalived/keepalived.conf file:
vrrp_sync_group VG1 {
group {
VI_1
}
}
vrrp_instance VI_1 {
state MASTER
interface eth0
virtual_router_id 51
priority 150
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
130.88.203.138
130.88.203.219
}
}
and the other has this:
vrrp_sync_group VG1 {
group {
VI_1
}
}
vrrp_instance VI_1 {
state BACKUP
interface eth0
virtual_router_id 51
priority 50
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
130.88.203.138
130.88.203.219
}
}
Your virtual_router_id on the backup needs to be incremented.
I don't agree with that. From keepalived.conf man page :
virtual_router_id is used to differentiate multiple instances of vrrpd
running on the same NIC
I'd rather replace NIC by network.
This means that, across a given network, each VRRP instance should have
its own/unique virtual_router_id.
But, as there is no "master VRRP instances" nor "backup VRRP instances"
but "virtual instances in master or backup state", all servers hosting
the same VRRP instance should have the same virtual_router_id for that
instance.
If you have a simple symmetric cluster, your conf file should be the
same on both servers, except for the following keywords in
vrrp_instance: state, priority, nopreempt and interface.
--
Sébastien BONNET -- Ingénieur système
Tel: 04.42.25.15.40 GSM: 06.64.44.58.98
|
