Eliminating duplicate entries from a list of machines in a partitioned database environment

If you are running multiple logical database partition servers on one computer, your db2nodes.cfg file contains multiple entries for that computer.

About this task

In this situation, the rah command needs to know whether you want the command to be executed only once on each computer or once for each logical database partition listed in the db2nodes.cfg file. Use the rah command to specify computers. Use the db2_all command to specify logical database partitions.

Note: On Linux® and UNIX operating systems, if you specify computers, rah normally eliminates duplicates from the computer list, with the following exception: if you specify logical database partitions, db2_all prepends the following assignment to your command:

export DB2NODE=nnn  (for Korn shell syntax)

where nnn is the database partition number taken from the corresponding line in the db2nodes.cfg file, so that the command is routed to the desired database partition server.

When specifying logical database partitions, you can restrict the list to include all logical database partitions except one, or specify only one using the <<-nnn< and <<+nnn< prefix sequences. You might want to do this if you want to run a command to catalog the database partition first, and when that has completed, run the same command at all other database partition servers, possibly in parallel. This is usually required when running the RESTART DATABASE command. You need to know the database partition number of the catalog partition to do this.

If you execute RESTART DATABASE using the rah command, duplicate entries are eliminated from the list of computers. However if you specify the " prefix, then duplicates are not eliminated, because it is assumed that use of the " prefix implies sending to each database partition server, rather than to each computer.