Upgrade Guide for DB2 UDB for z/OS > Database and UI Upgrade Planning > Planning Your Upgrade >

Planning Backup and Recovery Stages

You may need to recover all Siebel DB2 objects to a prior point in time. Your usual point-in-time recovery techniques may be insufficient in this environment.

In addition to the backup and recovery procedures that are standard for your environment, take a set of DB2 backups at key stages during the upgrade, using your preferred utility. A snapshot of your repositories and environment at these stages protects the progress of your upgrade in the event of a failed subsequent process.

It is recommended that you back up your repository at these key stages of the upgrade:

• Before any upgrade activity is started
• Before performing unloads

  NOTE:  Unloads must be performed when there is no system activity, so that the database is at a point of consistency.

• After upgrading the Siebel Database schema
• After the repository merge
• After upgrading the custom database schema

Review the results of all JCL jobs that you execute during the install or upgrade process. You can use a spool viewer such as IBM's SDSF to inspect the output from these jobs. You can review this information in addition to reviewing the upgrade log files.

CAUTION:  When performing a development environment, midtier-centric upgrade, if any of the preschm scripts do not run successfully, data is lost in S_ADDR_PER and cannot be recovered. To prevent this, you should specify that the upgrade script stops when it encounters an error so that you can correct it instead of letting the script run through and then dropping the temp table that contains the data. Use the -s option of the -stvf command to stop SQL command execution if an SQL error is returned. You should specify the -s option when running all Siebel upgrade scripts from the mid-tier.