Heartbeat detection

The name servers in a cluster use heartbeat detection to periodically verify that member servers are either accepting client connections or have failed. If a server is not accepting connections, the name server does not return profile (host:port) information to the client, and routes requests to other servers in the cluster. The name server also detects when a failed server is ready to accept connections again and starts routing client requests to that server.

If a name server using transient storage fails, the cluster rebinds automatically when you reboot the failed name server. Otherwise, the cluster provides access to components through the remaining name servers in the cluster.

If a name server using persistent storage and LDAP support fails, the cluster does not need to rebind, but LDAP may leave behind stale profiles, resulting in unnecessary client retries and failures. For this reason, Sybase recommends you use transient storage to support load balancing and high availability.

1. Select the Servers folder.

2. Highlight the name server for which you are configuring heartbeat detection.

3. Select File | Properties.

4. Open the Naming Service tab.

5. Select the Enable Heartbeat check box.

6. Enter the heartbeat frequency. This number is how often, in seconds, that the name server checks server availability. As the frequency period is shortened, server performance decreases. The default frequency is 120 seconds.

When you synchronize a cluster, the heartbeat settings (whether or not it is enabled and frequency) of the primary server are distributed to the other name servers in the cluster.