I'm having another difficult LVS / keepalived issue, now that I am trying
to add a backup/failover to the setup. With the config I am pasting below,
VRRP adverts are only sent out for my internal 10.0.0.1 VIP, none of the
externals, as I can see with tcpdump. Both servers seem to be stuck in the
master state, according to the logs, and `ip addr list` shows they are both
binding to all the VIPs.
Since keepalived doesn't die on a config syntax error, but silently
continues and behaves unexpectedly, perhaps I have made another formatting
or spelling error?
keepalived.conf:
global_defs {
router_id BBI_LVS1
}
vrrp_sync_group VG_1 {
group {
IN1
EXT1
EXT2
}
}
vrrp_instance IN1
{
state MASTER
interface eth1
virtual_router_id 51
priority 100
advert_int 1
lvs_sync_daemon_interface eth1
virtual_ipaddress {
10.0.0.1 dev eth1
}
authentication
{
auth_type PASS
auth_pass PqK8RyBB
}
}
vrrp_instance EXT1
{
state MASTER
interface eth0
virtual_router_id 52
priority 100
advert_int 1
lvs_sync_daemon_interface eth1
virtual_ipaddress {
206.114.148.54 dev eth0
}
authentication
{
auth_type PASS
auth_pass PqK8RyBB
}
}
vrrp_instance EXT2
{
state MASTER
interface eth0
virtual_router_id 53
priority 100
advert_int 1
lvs_sync_daemon_interface eth1
virtual_ipaddress {
206.114.148.55 dev eth0
}
authentication
{
auth_type PASS
auth_pass PqK8RyBB
}
}
virtual_server 206.114.148.54 22 {
delay_loop 10
lb_algo rr
lb_kind NAT
protocol TCP
real_server 10.0.0.100 22
{
weight 1
TCP_CHECK
{
connect_timeout 3
}
}
}
virtual_server 206.114.148.55 22 {
delay_loop 10
lb_algo rr
lb_kind NAT
protocol TCP
real_server 10.0.0.101 22
{
weight 1
TCP_CHECK
{
connect_timeout 3
}
}
}
virtual_server 206.114.148.54 80 {
delay_loop 10
lb_algo rr
lb_kind NAT
protocol TCP
real_server 10.0.0.100 80
{
weight 1
TCP_CHECK
{
connect_timeout 3
}
}
real_server 10.0.0.101 80
{
weight 1
TCP_CHECK
{
connect_timeout 3
}
}
}
|