Debian Jessie Keepalived simple Redis Cluster

Even though keepalived under Debian Jessi give samples of check local service with killall in /usr/share/doc/keepalived/samples/keepalived.conf.vrrp.localcheck

vrrp_script chk_haproxy {
       script "killall -0 haproxy"     # cheaper than pidof
       interval 2                      # check every 2 seconds
}

But Jessie only have SystemD. [killall] has been remove with SystemV.
You can still use killall5 which is the SystemV killall command.

But killall5 simulates the killall command of SysV UNIX; it sends the given signal to all processes owned by the calling UID. When does it root, it can do very bad things. BE VERY CAREFUL with the killall command on!


# cat /etc/keepalived/keepalived.conf

vrrp_script chk_redis {
        script "pidof redis-server"     #  pidof inside of killall, just to be safe.
        interval 2                      # check every 2 seconds
        weight 2                        # add 2 points of prio if OK
}

vrrp_instance VI_1 {
        interface eth0
        state MASTER
        virtual_router_id 59
        priority 100                    # 101 on master, 100 on backup
        virtual_ipaddress {
            192.168.1.10
        }
        track_script {
            chk_redis
        }
}

In my opinion pidof is safer than killall or killall5.