Increased data availability

When you distribute table and index fragments across different disks or devices, you improve the availability of data during disk or device failures. The database server continues to allow access to fragments stored on disks or devices that remain operational.

This availability has important implications for the following types of applications:
  • Applications that do not require access to unavailable fragments

    A query that does not require the database server to access data in an unavailable fragment can still successfully retrieve data from fragments that are available. For example, if the distribution expression uses a single column, the database server can determine if a row is contained in a fragment without accessing the fragment. If the query accesses only rows that are contained in available fragments, a query can succeed even when some of the data in the table is unavailable. For more information, see Designing an expression-based distribution scheme.

  • Applications that accept the unavailability of data

    Some applications might be designed in such a way that they can accept the unavailability of data in a fragment and require the ability to retrieve the data that is available. To specify which fragments can be skipped, these applications can execute the SET DATASKIP statement before they execute a query. Alternatively, the database server administrator can use the onspaces -f option to specify which fragments are unavailable.

If your fragmentation goal is increased availability of data, fragment both table rows and index keys so that if a disk drive fails, some of the data is still available. If applications must always be able to access a subset of your data, keep those rows together in the same mirrored dbspace.


Copyright© 2018 HCL Technologies Limited