OpenSwitch supports a redundant environment with two “mutually-aware” OpenSwitch servers that serve the same pools of two Adaptive Servers. Each OpenSwitch server is aware of the other OpenSwitch server and whether the Adaptive Servers for which that server is responsible are ready to accept client connections.
For example, let’s say you have one OpenSwitch server (OSW1) connected to one Adaptive Server (Server1). You have another OpenSwitch server (OSW2) and another Adaptive Server (Server2) set up for failovers. OSW1 fails and its connections to Server1 fail over to OSW2. Then Server1 fails over to Server2. In versions earlier than 15.0, when you restarted OSW1, it would continually try to reconnect to Server1 because it had no knowledge that Server1 was not running.
In a mutually-aware OpenSwitch implementation, when OSW1 restarts, it queries the companion OpenSwitch server (OSW2), which tells it that Server1 has failed, therefore, it does not try to reconnect to the unavailable server.
To ensure that this setup works properly, both OpenSwitch servers must:
Have the same knowledge about each pool’s and server’s status at any given time.
Store and retrieve the redundancy information from a set of locations that are accessible to all components in the OpenSwitch implementation to ensure that the effective status persists between OpenSwitch restarts.
Before you install and configure mutually-aware OpenSwitch servers, be sure to read this chapter and make sure that you review the mutually-aware OpenSwitch server “Requirements”.
| Copyright © 2005. Sybase Inc. All rights reserved. |
|
|
