306 Solving Operational Business Intelligence with InfoSphere Warehouse Advanced Edition
node to reference it. To create a single configuration file that is shared across all
nodes on which Optim High Performance Unload is installed:
1. Make a copy of the default configuration file, customize it as required, and
save it on a file system that is accessible across all nodes; using /db2home is
recommended.
2. Add the dir_cfg parameter to the default Optim High Performance Unload
configuration file on each data node where Optim High Performance Unload
is installed to reference the shared Optim High Performance Unload
configuration file.
3. Make further customizations to the referenced configuration file only.
Create a directory structure for Optim High Performance
Unload staging and output files
A table in a table space that is 100 GB in size requires up to 100 GB of disk
space for staging data. Additional disk space is required for staging data if
incremental backup images are referenced.
Staging area
You can specify only one directory for the staging area. Create the same
directory structure on each data node to accommodate staging of data. Where
possible, locate the output and staging directories for each node on a separate
file system on separate logical unit numbers (LUNs). Size the staging area to
accommodate the largest table space in your database, and also take into
account the incremental and delta backup policy.
Output area
To enable parallelism when using Optim High Performance Unload, create an
output directory for each database partition. Size the output area to
accommodate the largest table space in your database. Avoid using the same
file systems as used by your database; you do not want DB2 to run out of space
on the production database.
Set configuration file parameters
The product configuration file holds default parameter values that are referenced
each time the Optim High Performance Unload utility is used. Recommendations
apply to these parameters:
???? nbcpu
Determine the number of work threads to be initiated by assessing the
recovery time objective required and the resources available. Change the
configuration parameter as required before running the db2hpu command.