Partition.pmd

CDB/RePart61CDB/RePartIntroductionMultiple-dataset tablespaces are the future of DB2. They offer significant ad-vantages by allowing concurrent and independent application and utility pro-cessing of the different datasets. Very large databases, used concurrently bylarge amounts of users, require the data to be placed in many small “vessels” asopposed to a single, large container.Multi-dataset objects, however, have a down side, and that is complexity. DB2supports, at the time of this writing, 254 tablespace partitions, which require inturn the same number of clustering index partitions. Tablespaces of this sizemay have non-partitioning indexes, which are likely to be in multiple pieces ordatasets. Therefore, today’s large tablespace may force you to deal with over500 objects, which may be split into 1000 datasets or so. The Data Base Ad-ministrator in your organization must not only define these massive objects,but also manage them over time, and keep them running at optimal perfor-mance (which is why the tablespace was partitioned to start with)CDB/RePart provides multiple tools to help you manage these objects.Among other functions, it lets you:• Rebalance partitioned tablespaces so that they provide opti-mal performance at all times.• Convert your fast-growing tables to partitioned.• Repartition partitioned tablespaces (that is, change the num-ber of partitions).• Manage “periodic data,” that is, business data that is gener-ated, used ...