An active-passive configuration is a configuration with two or more nodes and a single Adaptive Server. The set of nodes that hosts the Adaptive Server under normal conditions is called the primary nodes; the set of nodes that can potentially host the Adaptive Server is called the secondary nodes.
When the Adaptive Server or any of the resources it depends on, such as a disk or the node itself, crashes, the Adaptive Server, along with all required resources, is relocated and restarted on a secondary node. This movement from the primary node to the secondary node is called failover.
After failover, the node hosting Adaptive Server is considered the primary node until the System Administrator performs a planned failback, or until the Adaptive Server on the new primary node fails, causing another fail over.
After failover, all existing client connections are lost.
The clients must reestablish their connections and resubmit any
uncommitted transactions as soon as the Adaptive Server is started
on the secondary node. The client connection failover can be performed
automatically by using high availability connections and self-referencing
the hafailover entry in the interfaces file. See “Configuring the interfaces file on the client side” for information.
You can configure the active-passive setup with multiple secondary nodes so that Adaptive Server can survive multiple failures. With a multi-node setup, Adaptive Server can service requests from clients as long as at least one of the primary and secondary nodes is available to host the Adaptive Server and its resources. See “Working with a multi-node cluster” for more information.
| Copyright © 2005. Sybase Inc. All rights reserved. |
|
|
