Re: need help with vrrp. using keepalived LVSDR

[Janusz Krzysztofik <jkrzyszt@xxxxxxxxxxxx>]

Tommy Butler napisał(a):
> Two LVS directors, one master, one for failover, and 4 realservers behind
> them.  LVS-DR.  The LVS part works fine.  The failover does not.
>
> Primary configuration that is problematic appears below:
>
> global_defs {
>   notification_email {
>     tommy@xxxxxxxxxxxx
>   }
>   notification_email_from tommy@lvs1
>   smtp_server localhost
>   smtp_connect_timeout 30
>   router_id LVS_DIRECTOR_1
> }
>
> vrrp_sync_group VG1 {
>   group {
>      VI_1
>      VI_2
>   }
> }
I think you don't need any group, see below.
> vrrp_instance VI_1 {
>    interface eth0:0
Just use eth0 here
>    state MASTER
>    lvs_sync_daemon_inteface eth0:0
and here
>    virtual_router_id 50
>    nopreempt
>    priority 150
>    advert_int 1
>    smtp_alert
>    virtual_ipaddress {
>        64.147.XXX.70
put both VIPs here
>    }
> }
and remove this second instance.
> vrrp_instance VI_2 {
>    interface eth0:1
>    state MASTER
>    lvs_sync_daemon_inteface eth0:1
>    virtual_router_id 51
>    nopreempt
>    priority 150
>    advert_int 1
>    smtp_alert
>    virtual_ipaddress {
>        64.147.XXX.72
>    }
>    authentication {
>       auth_type PASS
>       auth_pass LVSPASS
>    }
> }
I have never tried configuring keepalived with old-style interface 
aliases but I suspect it would not work as expected. Have you checked 
for any related messages in your syslog?
Janusz