High-Availability Cluster Environments
Consider the following information when planning to use Connect:Direct® for UNIX in a high-availability cluster environment.
Supported High-Availability Cluster Environments
Connect:Direct for UNIX is certified to operate in the following high-availability cluster environments:
- IBM® high-availability cluster multiprocessing (HACMP) environment
- Hewlett-Packard MC/Service Guard
- SunCluster versions 2.2, 3.0, and 3.2
- Veritas Infoscale Availability (formerly Veritas Cluster Server)
If you plan to install Connect:Direct for UNIX in a high-availability cluster environment, complete the following tasks:
- Install the clustering software on each computer in the cluster, including setting up a logical host or application package.
- Create a user with the same name and user ID on each cluster node.
- Create a Connect:Direct subdirectory on a shared file system on a shared volume group.
- Ensure that the shared file system is owned by the IBM Connect:Direct user.
- Install IBM Connect:Direct on the shared file system.
- Perform the procedures necessary to define the high-availability settings and configure the cluster environment.
Limitations of High-Availability Clusters
When running Connect:Direct for UNIX in a high-availability cluster environment, be aware of the following limitations:
- If a failure occurs, all Processes being held will be restarted when IBM Connect:Direct is restarted. This includes Processes that are held by the operator as well as Processes held in error. This could cause a security risk.
- When a IBM Connect:Direct ndmsmgr process associated with a IBM Connect:Direct Process is killed, the Process is not automatically restarted and is put in the Held in Error state. It must be manually restarted; otherwise, the IBM Connect:Direct Process is restarted when the cluster restart occurs.