Solving Operational Business Intelligence with InfoSphere Warehouse Advanced Edition

304 Solving Operational Business Intelligence with InfoSphere Warehouse Advanced Edition

򐂰 TAKEN AT

Use this parameter to specify a partial backup image time stamp to reference

multiple backup images, including incremental and delta backup images.

For example, a time stamp of 20120630 instructs the db2hpu command to

look for all backup images on June 30, 2012. Where multiple full backup

images contain the partial time stamp, Optim High Performance Unload

terminates and outputs a message to that multiple backup images with the

same partial time stamp were found.

To allow Optim High Performance Unload to determine the most recent

backup images to use, omit the TAKEN AT clause.

򐂰 ON SOURCE HOST

Use this parameter to instruct Optim High Performance Unload to unload data

in parallel to individual output files on each database partition instead of

creating a single large file on the administration node.

This action increases parallelism in a shared-nothing architecture because

each output file is created locally on each data node. Each data node

involved recovers the data for its local DB2 database partitions to its local

output path with the correct file per node specified by the OUTFILE key.

򐂰 OUTFILE with {source_host} and {source_node} keywords

Use the OUTFILE("%{source_host}:/OUTPUTPATH/FILE.%{source_node}")

keywords to generate the output file on the correct file system associated with

each database partition.

Each database partition has a separate output file created with the database

partition number appended to the name of the file.

8.4.6 Create the control file and issue a db2hpu command

Follow these recommendations when creating a control file for an unload

operation:

򐂰 Specify the data to be recovered in the control file by using the WHERE

clause to specify the range of values from columns; using the

DATAPARTITION ID to reference particular table range; and using the PART

clause to reference particular database partitions.

򐂰 Reference a database catalog backup image that is consistent with the data

to be restored. For example, if you are unloading data for a dropped table you

need to reference a database catalog that has the definition of that dropped

table. Using an up-to-date catalog reduces the need for Optim High

Performance Unload to access the backup image through tape.