For HA-cluster I would look into a two-node active/passive cluster using
Heartbeat as the cluster manager and
DRBD to replicate the filesystems realtime across the two cluster nodes.
I've been managing clusters built this way for four years now - all on Redhat since it is policy to use that distro, but I built the RPM's for heartbeat and drbd myself, and it would be fairly trivial to do the same for Slackware packages. You will have a rock-solid cluster that way, and except for catastrophic failure in your server room, your ens users would perceive a near 100% availability of their services (I run Samba, LDAP, IMAP, MySQL, Postfix, Apache and more as HA services - typical medium-sized office services)
As for LVS, their project at
http://www.linuxvirtualserver.org/ works perfectly well on a Linux cluster used as loadbalancer with (again) Heartbeat as the cluster manager. I setup such an environment once as proof-of-concept for a client with a Linux Terminal Server farm (Nomachine/NX based) behind the load balancer. Great fun that was.
Eric