Space announcement

   

This space provides the same content as before, but the organization of the home page has changed. The content is now organized based on logical branches instead of legacy book titles. We hope that the new structure will help you quickly find the content that you need.

How SHRLEVEL CHANGE differs in REORG PLUS and IBM Db2 REORG

REORG PLUS has several features that distinguish it from the IBM Db2 REORG utility. These features improve performance, make a SHRLEVEL CHANGE reorganization easier to administer, and provide the least disruptive reorganization to applications that are accessing the data.

SHRLEVEL CHANGE processing differences from Db2 REORG

The SHRLEVEL CHANGE option of REORG PLUS allows Db2 table space and index space objects to be available for RW access by Db2 applications during most of the reorganization.

Allocating storage versus using Db2 resources

To be truly nondisruptive, a reorganization utility must do more than simply allow application access. If the reorganization utility competes with the application for database resources, disruption in the form of degraded performance can still occur. To avoid this situation, REORG PLUS does not use the Db2 buffer pool, buffer manager, data manager, or work databases to accomplish its tasks.

Instead, REORG PLUS allocates data sets, staging copies of the Db2 objects, and virtual storage as needed. This additional DASD and utilization of virtual storage space eliminates the use of Db2 resources to accomplish the reorganization. In addition, the RID map is not a Db2 object in REORG PLUS as it is in the IBM Db2 REORG utility. This factor prevents RID map access from interfering with Db2 processing. By eliminating contention for these resources between REORG PLUS and the application, the reorganization can proceed in a less disruptive manner.

Determining the longlog condition

REORG PLUS and Db2 REORG differ fundamentally in how they determine a longlog condition. Db2 REORG has an internal buffer that contains log records. When this buffer becomes full, Db2 REORG processes the records. Db2 REORG then repeats the process, making another pass at the next batch of log records. It uses an average of the last n passes to determine whether a longlog condition exists, or whether processing can be completed by the time that is set by the DEADLINE option or by the MAXRO value.

In contrast, REORG PLUS samples the arrival rate of the log records every 15 seconds to determine the rate at which the records are being applied. It uses a weighted average of the last 15 samples to determine whether a longlog condition exists, or whether processing can be completed by the time that is set by DEADLINE or by the MAXRO value.

By using more samples and a weighted average, REORG PLUS provides the following advantages over Db2 REORG:

  • A greater level of accuracy in determining whether processing can be completed by the time that is set by DEADLINE or by the MAXRO value

  • An increased probability that sudden changes in activity levels will not force unnecessary longlog conditions

  • The ability to specify LOGTHRSHLD to define when the process will end

Online partition rebalancing

When you use a DDLIN data set with ALTER statements, REORG PLUS performs online partition rebalancing and updates limit keys for you with no outage and without your space entering REORP status. Online partition rebalancing is performed on the staging data sets and the limit-key updates are made in the UTILTERM phase. For more information, see DDLIN data set restrictions and considerations.

Alternatively, you can use the REBALANCE command option to rebalance partitions. By specifying REBALANCE, you can have REORG PLUS determine the limit keys for you. For more information, see the REBALANCE option description.

SHRLEVEL CHANGE administrative differences from Db2 REORG

To reduce the administrative burden of running an online reorganization, REORG PLUS has automated and simplified some of the tasks.

RID translation maps

One of the memory structures that is required for a log apply reorganization is a RID translation map, which REORG PLUS maintains in data spaces. This object is required only during the reorganization so there is no need for it to persist after the reorganization is complete. The RID translation map is a temporary object.

REORG PLUS performs the following RID map tasks:

  • Allocates it for you

    You do not need to define a data set or Db2 table in advance to contain the RID map.

  • Assigns the name and ensures that it is unique across your systems

    There is no contention among reorganization jobs for the use of a RID map of a particular name.

Staging data set allocation

REORG PLUS offers several options for defining the staging data sets. You can choose the option that best suits your environment:

  • Define the staging data sets through IDCAMS commands that you supply to REORG PLUS in the SYSIDCIN DD statement for VCAT-defined objects.

  • Have REORG PLUS automatically define the staging data sets for storage-group-defined objects.

  • Predefine the staging data sets by using a separate IDCAMS step, and specify NO on the REDEFINE command or installation option.

You can instruct REORG PLUS to complete one of the following actions after the reorganization is complete:

  • Delete the old data sets that previously held your Db2 table space and index space objects

  • Automatically rename the data sets to the staging data set names

Related topic

Was this page helpful? Yes No Submitting... Thank you

Comments