Network Router User Manual

88 GINA V4.0 System Administrator Guide – September 2000
Revision generation
6.3 Revision generation
The purpose of the revision generation is as follows: when the configuration is revised, the
generation for the applications not affected by the revision remain the same.
Prerequisites
The use of a repository is an absolute must for this revision generation (see the statement
REPOSITORY on page 60). The repository must be created when generating the previous
version or updated to the new status.
The applications use host resources such as authorization keys (KEYVECTOR) and port num-
bers (PORTADDRESSES). The applications not affected by a revision must use the same
resources. The authorization keys and port numbers are stored in the repository for each
application.
When creating the kdcdf script, the generator config must generate unique identifiers
for each application. The same identifiers must then be used in a revision generation. In the
case of a first generation with an empty repository, a generation number is written to the
repository for each application so that this number will be used to generate the identifiers
(see the statement
START_VALUE on page 64).
The following further characteristics or values are entered in the repository for each server
application:
– KDCFILE, DOUBLE
If the KDCFILE is maintained in duplicate for security reasons, the characteristic
DOUBLE is stored and can then be taken into account during the next generation.
– PGPOOLFS
The PGPOOLFS parameter defines the number of files across which the page pool is
distributed. This number is entered in the repository and can thus be evaluated during
the next generation.
– RECBUFFS
RECBUFFS defines the number of files across which the restart area is to be distributed.
This number is stored in the repository and read during the next generation.