The logical
cluster is an abstract representation of one or more instances in
a physical shared-disk cluster used to manage workload,
failover, and client application access to the cluster.
Each logical
cluster has a set of instances on which it runs, and can have a set of
instances to which it fails over. Routing rules direct incoming
connections to specific logical clusters based on an application,
user login, or server alias supplied by the client. Other rules
can restrict the logical cluster to a bound connection or allow access by any
authenticated connection.
- In the left pane of the Administration Console, expand , then select Logical Cluster.
- Click the drop-down arrow and select New.
You see the Add Logical Cluster wizard.
- Select the shared cluster to which to add the logical cluster.
- Assign a name to the logical cluster that is representative
of the job it will perform, such as SalesLC.
- Click Add to select the instances you want to add to the logical cluster. These instances are considered "base" instances.
You can remove an instance from this list by clicking Remove.
- In the Failovers screen, select Add to choose the failover server instances on which the logical cluster will run if one or more of the base
instances fail. Any instance in the physical cluster can be a failover
resource.
- Select the group number for the failover instances. The failover group lets
you specify preference and order in which failover instances
are used. Lower-numbered groups are used first.
- Click Add to add routed applications, logins, or aliases.
Routing
rules allow you to specify specific
applications, logins, and aliases to which the logical cluster will connect. See the
Cluster Users
Guide for more information about routing.
Add additional
routes by selecting:
-
Application Route – specify the name of the application in the
Add Application Route screen.
- Alias Route – specify
the alias name in the Add Alias Route screen.
- Login Route – select
the login name from the list in the Add Login Route screen (hold
the Ctrl key to select multiple names).
You can drop a listed route by selecting the route name and clicking Remove.
- On the Load Profile screen, click Change to select a load profile to assign to the logical cluster.
The Cluster Edition offers two system load profiles: sybase_profile_oltp for OLTP environments and sybase_profile_dss for DSS environments. You cannot modify or delete system load profiles.
- On the Set Options page:
- Select Cluster or Instance for the system view.
The system view determines whether the logical cluster users view the Cluster
Edition as an entire cluster or as individual instances. This
affects some queries and stored procedures.
- (Optional) Select Automatically start logical cluster to start the logical cluster when the shared cluster is started.
- Choose instance or group to specify whether the instances
are brought online as a group or individually.
- (Optional) Select Fail to any to specify whether any instance, or a specific instance
can be a failover resource.
- Choose the down-routing mode to specify how client
connections are routed if the logical cluster designated by the
routing rule is not available.
-
system – sends
unroutable connections to the system logical cluster.
-
open – sends unroutable
connections to the open logical cluster.
-
disconnect – disconnects
unroutable connections.
- Choose a login distribution mode to specify how connections are distributed in logical clusters with multiple instances.
Note: The login distribution mode does not affect single-instance logical clusters.
- affinity – specifies that a connection accepted by an instance is retained as long as the target logical cluster is running on that instance.
If the load profile specifies a load threshold, and the load on the instance is too high, the workload manager redirects the connection to the least loaded instance in a logical cluster.
If the target logical cluster is not running on the instance, the workload manager redirects the connection to the least loaded instance in a logical cluster.
- round-robin – specifies that incoming connections are distributed in a round-robin fashion among the instances hosting a logical cluster. For example, if “SalesLC” is running on “ase1” and “ase2”, the workload manager sends the first connection to “ase1,” the next to “ase2”, and so on. Load scores are not included in the algorithm.
- (Optional) Select Open role to indicate that all connections that are not routed to a logical cluster via an explicit
routing rule are routed to the current open logical cluster. When
you create a new cluster, the system logical cluster is automatically
designated the open logical cluster. You can reassign the open role
to another logical cluster. However, only one open logical cluster
can exist per physical cluster.
- (Optional) Click Summary to review your information.
- Select Finish to build the logical cluster.