1.1.1

Table Of Contents
Connect to a Distributed System Using Locators
SQLFire thin clients specify the client bind address and port of a locator in order to connect to a SQLFire cluster.
Connecting to a locator ensures that a client is load-balanced to an available SQLFire server in the distributed
system. If the client's connection to a server is dropped (for example, if the server is shut down), then the client
can retry connecting to the same locator to access another available server.
Note: Thin clients do not specify the peer discovery addresses or port as do SQLFire servers and peers.
For example, in an interactive sqlf session, use the connect client command and specify the client host
and port number of a locator:
connect client 'locator1host:1527';
In a third-party JDBC tool, specify the host and port in the JDBC connection string:
jdbc:sqlfire://locator1host:1527
Locators balance client requests to all SQLFire servers available in a cluster, so multiple connection requests to
the same locator may result in connections to different servers.
Configure Multicast Discovery
You congure multicast discovery by specifying the mcast-address and mcast-port properties.
To start an embedded SQLFire peer with multicast discovery, specify the multicast address and port as options
in the JDBC peer client connection URL, or in the property object that you send in the FabricServer.start()
method. For example:
jdbc:sqlfire:;mcast-address=address;mcast-port=port
To start a SQLFire server with multicast discovery from a command prompt or terminal window, add the multicast
properties directly to the sqlf command. For example:
sqlf server start -mcast-address=address -mcast-port=port
Note: Thin JDBC clients and ADO.NET clients do not connect to a cluster using multicast. Instead, they
must specify the address and port of a single peer in the cluster.
225
Configuring Discovery Mechanisms