Remember the old days when DB2 had to handle only a couple of objects? You may recall:
- A quiet and peaceful batch window` with exclusive access to the data
- No application causing annoying database maintenance
- Consistent SHRLEVEL NONE COPIES; REORG, and RUNSTATS without obstructing locks.
In today’s on-demand world, nothing seems to stop. Everything must be just-in-time, with continuous availability, excellent performance, and ASAP!
To assure on-demand business services, each underlying process must support on-demand requirements. If you run e-business applications, Enterprise Resource Planning (ERP), Customer Relationship management (CRM) or Java applications, you need an appropriate infrastructure.
You’re running a mainframe, perfect! IBM is committed to system availability. You run DB2 V8 or even DB2 9? Great, you’re already equipped to face the challenge. Now you must react nimbly to only upcoming requirements. DB2 for z/OS handles more than 9,000 transactions a second. So what does a database maintenance solution that can meet today’s business demands look like?
The Five Levels of Database Maintenance
Database maintenance can be characterized on different levels from one to five (see Figure 1):
- Cyclic database maintenance
- Reactive maintenance
- Automatic maintenance
- on-demand database maintenance
- Autonomic database maintenance.
At level one, cyclic database maintenance, a DBA writes a utility job for all objects of the database. This procedure occurs periodically.
At level two, reactive maintenance, the procedure in level one has become too time-consuming. only when an object required maintenance was a utility job built and executed. This reduced the number of utility jobs as well as run-time, CPU and I/O, but it’s a reactive approach. An upcoming performance bottleneck can lead to an outage even before a DBA learns about it.
At level three, automatic maintenance, if there’s a maintenance requirement, a utility is automatically generated and executed. Within a pre-defined, static job net, each single object is frequently checked for maintenance requirements based on defined thresholds. At this level, granular automation comes up with the database maintenance process. Utility job generation is included; no one has to start or execute the process steps. The entire procedure is defined in the company’s job scheduler. The job net definition typically covers four major steps: