DB2 Configuration Examples

A few examples of what happens during a failover using LifeKeeper for Linux DB2 Recovery Kit are provided below. In the following pictures, EE and EEE are used to denote database configurations; ESE may be substituted wherever appropriate.

Configuration 1: DB2 Single Partition Active/Standby Configuration

The DB2 instance is protected on Server 1. Server 2 will assume the DB2 resources when a failure occurs.

Configuration 2: DB2 Single Partition Active/Active Configuration

One DB2 instance is protected on Server 1 and another DB2 instance is protected on Server 2. Each server will assume the other’s resources when a failure occurs.

Configuration 3: DB2 Multiple Partition Active/Standby (1 Cluster)

One DB2 instance with two database partition servers is protected on Server 1 with one LifeKeeper DB2 resource hierarchy. Server 2 will assume ownership of the DB2 resource hierarchy when a failure occurs.

Note: For all cluster of cluster configurations listed in the following section, users should be aware that the cluster of cluster configuration is protecting only one DB2 instance with multiple partitions on multiple physical nodes.

Configuration 4: DB2 Multiple Partition Active/Standby (Cluster of Clusters)

One DB2 instance with two database partition servers is protected on Server 1 and two database partition servers protected on Server 3. There is one LifeKeeper DB2 resource hierarchy on Server 1, extended to Server 2, and another DB2 resource hierarchy on Server 3 extended to Server 4. When a failure occurs on Server 1, Server 2 will assume its resource. When a failure occurs on Server 3, Server 4 will assume its resource.

If the server that is exporting the DB2 instance home directory and its backup server become inoperable at once, the DB2 database is inaccessible. In addition, if the NFS hierarchy for the exported DB2 instance directory (primary and all backups) become inoperable at the same time, the DB2 database will be inaccessible until the NFS hierarchy can be restored.

Configuration 5: DB2 Multiple Partitions Active/Active (1 Cluster)

One DB2 instance with one database partition server is protected on Server 1 and one database partition server protected on Server 2. There is one LifeKeeper DB2 resource hierarchy on Server 1 and another DB2 resource hierarchy on Server 2. When a failure occurs each server will assume the other’s resources.

Configuration 6: DB2 Multiple Partitions Active/Active (Cluster of Clusters)

One DB2 instance with two database partition servers is protected on Server 1, one database partition server protected on Server 2, one database partition server protected on Server 3 and two database partition servers protected on Server 4. There is one LifeKeeper DB2 resource hierarchy on each server in the cluster. Upon failure, Server 1 and Server 2 assume each other’s resources and Server 3 and Server 4 assume each other’s resources.

If the server that is exporting the DB2 instance home directory and its backup server become inoperable at once, the DB2 database is inaccessible. In addition, if the NFS hierarchy for the exported DB2 instance directory (primary and all backups), become inoperable at the same time, the DB2 database will be inaccessible until the NFS hierarchy can be restored.

Configuration 7: DB2 Multiple Partition (4 Node Fibre Channel Cluster)

One DB2 instance with two database partition servers is protected on Server 1, one database partition server protected on Server 2, one database partition server protected on Server 3 and two database partition servers protected on Server 4. There is one LifeKeeper DB2 resource hierarchy on each server in the cluster. Each server in the cluster provides backup protection for the other in the event of failure.

If the server that is exporting the DB2 instance home directory and its backup server become inoperable at once, the DB2 database is inaccessible. In addition, if the NFS hierarchy for the exported DB2 instance directory (primary and all backups), become inoperable at the same time, the DB2 database will be inaccessible until the NFS hierarchy can be restored.