<!-- ""@(#)loghost 1.1 97/10/03 Sun Microsystems Inc"" >
 
<title>Logical Host</title>
<h1>Logical Host</h1>

A data service is made highly available by exploiting the concept of a
logical host. The data  service's  data is placed  on the  disk groups
associated with a  logical host. A  logical host's disk  groups may be
used  for raw data  storage or for  filesystems or  both. The physical
disks associated  with the disk groups may  be in storage devices that
are connected to multiple  nodes. The data  service makes its services
accessible to clients on the network by advertising well known logical
addresses associated with the logical host.  The logical addresses are
part  of the IP  name space at a site,   but do not  have any physical
systems associated   with them. The  clients  should use these logical
addresses to  access the  services provided  by  the data  service.  A
logical  network IP address failover occurs  as part of a logical host
failover, causing network    clients of the  data service   to recover
quickly, and in some cases, transparently.<p>

A data service  can   be registered on   single or   multiple  logical
hosts. It is easier to design, implement, and  maintain a data service
that uses just one logical host. In this  case, all the data service's
data  is  placed on just that   logical host's  disk  groups. The data
service might need just one set of server  process(es) or daemon(s). A
physical host runs the  daemon(s) for  that data  service only  if the
physical host currently masters the single logical  host that the data
service uses. When the physical host takes over mastery of the logical
host,  the data service's   start   method registered with the    Data
Services API  can start  up the daemon(s).  When  the physical host is
giving up mastery of the logical  host, the data service's stop method
registered with the Data Services API can stop the daemon(s).<p>

If multiple logical  hosts are used, the data  service's data must  be
split into disjoint sets. The sets must  be split so that no operation
the data  service needs  to perform  requires data  from more than one
set.  Using multiple logical   hosts   enables some rudimentary   load
balancing  - different   logical hosts  can  be  mastered on different
physical  hosts,  and  each physical host   handles   the data service
trafffic only  for that logical host  it  masters. Thus,  all physical
hosts in  the cluster are doing useful  work in addition to serving as
backups for each other.<p>
