今度はkeepalived(1.1.15)でもたついている

repcached(memcached+レプリケーション)を２台のサーバで起動して、マルチマスターで動くことまでは確認できた→(テスト進行表)
そこで、keepalivedでロードバランスと、health checkをしようとして、keepalivedのインストール、起動までは問題が無かった。しかし、設定がちゃんと反映されているかを確かめたくて、/etc/sysconfig/keepalived に -D -d のコマンドオプションを指定して service keepalived startしたらば、下記のようにGlobal definitionsしか表示されない。(環境はFedora5 カーネルは2.6.15-1.2054_FC5でx86_64)

today now servername Keepalived: Starting Keepalived v1.1.15 (05/10,2008)
today now servername Keepalived: Starting VRRP child process, pid=25759
today now servername Keepalived_vrrp: Using MII-BMSR NIC polling thread...
today now servername Keepalived_vrrp: Registering Kernel netlink reflector
today now servername Keepalived_vrrp: Registering Kernel netlink command chann
el
today now servername Keepalived_vrrp: Registering gratutious ARP shared channe
l
today now servername Keepalived_vrrp: Opening file '/etc/keepalived/keepalived
.conf'.
today now servername Keepalived_vrrp: Configuration is using : 59219 Bytes
today now servername Keepalived_vrrp: ------< Global definitions >------
today now servername Keepalived_vrrp:  Router ID = LVS_MEMCACHED
today now servername Keepalived_vrrp:  Smtp server = 127.0.0.1
today now servername Keepalived_vrrp:  Smtp server connection timeout = 30
today now servername Keepalived_vrrp:  Email notification from = root@hostname
today now servername Keepalived_vrrp:  Email notification = mgrsite@hostname

ちなみに/etc/keepalived/keepalived.confはこんな感じ

global_defs {
   notification_email {
     mgrsite@amuse.co.jp
   }
   notification_email_from root@mysql1.amuse.co.jp
   smtp_server 127.0.0.1
   smtp_connect_timeout 30
   router_id LVS_MEMCACHED
}


virtual_server_group MEMCACHED {
    192.168.1.10 11213
}
virtual_server MEMCACHED {
    delay_loop 6
    lb_algo rr
    lb_kind NAT
    persistence_timeout 50
    protocol TCP

    real_server 192.168.1.10 11211 {
        weight 1
        TCP_CHECK {
                connect_port 11211
                connect_timeout 3
        }
    }
    real_server 192.168.1.11 11211 {
        weight 1
        TCP_CHECK {
                connect_port 11211
                connect_timeout 3
        }
    }
}

あー、もしかしてkeepalivedって、ルータ構成（NIC２枚挿しで、loadbalancerとreal_serverは別サーバ）というのを仮定してるのか？つまり、load balancerとreal_serverの１台が同じマシンじゃダメとか、NICが１枚じゃダメとか...
（追記）
別のマシンにkeepalivedをインストールしてみたが、特に挙動は変化なし。Keepalived for LVSのUserguideの8.3節の２のようなログを期待してるんだけど....