Kbase P4286: Is the Progress Enterprise Database suitable to run on an IBM AIX HACMP-cluster?
| Autor |
Progress Software Corporation - Progress |
| Acesso |
Público |
| Publicação |
15/05/2008 |
|
Status: Verified
GOAL:
Is the Progress Enterprise Database suitable to run on an IBM AIX HACMP-cluster?
GOAL:
Does Progress support clustering?
FACT(s) (Environment):
IBM AIX
Progress 9.x
OpenEdge 10.x
FIX:
Progress does not support the **same** database running in multiple machines or nodes in a cluster environment, this configuration is used for performance reasons. For cluster failover purpose Progress offers a product called Fathom High Availability Cluster (included in Enterprise Database Licenses since 9.1E).
Fathom High Availability Cluster works together with IBM HACMP technology and provides an automated way to move the database resources from another node in case of a failure on the primary node.
This is an example of a failover configuration: if node 1 fails, a backup node is in place that will use the database crash recovery mechanism to restart all database functions. As long as the failover processes, caching etc. are transparent for the database engine, then there should be no issues.
The only thing HACMP does is make sure that the correct disks are mounted and the system gets the right IP-address. The startup and shutdown scripts are equal on both local disks. The database disks are **shared** between the production- and stand-by box.
Without Fathom High Availability Cluster product, the user has the option to configure scripts to control the failover event and move the database resources to another machine. It is very important to mention that these scripts have to be tailored manually by the System Administrator and are not supported by Progress. On the other hand, Progress will also run as usual without enabling Failover Clusters in a clustered environment.