It is very important to plan the overall implementation of your archiving infrastructure from the beginning. If you intend to limit the whole implementation to be completely within a single computer, then the planning is straightforward since there is a network of one computer, which is clearly the simplest model. Other enterprises have many computers, so must consider:
The databases you intend to extract data from and the computers where these databases run. Data Archiving for Adabas must be installed on all these computers in order to run extractors.
The computers you intend to run accumulators on, to write archive data. Accumulators can be run on different computers to extractors, so Data Archiving for Adabas must be installed on all the accumulator computers to:
spread processing load and
run accumulators on cheaper hardware.
Once you have an overview of all the computers involved, you now must decide where the Adabas file used for storing the configuration data is to be loaded and run. Software AG strongly recommends a single file is shared by all computers. To accommodate file sharing, we have provided an in-built mechanism - so you do not need to acquire our Entire Net-Work product simply for configuration. By sharing the file, all connectivity configuration is automatic, which completely frees you from the chores of reciprocally matching up numerous settings across all computers. This automated configuration is a major feature.
The file usually runs on one of the primary computers out of all involved, but the choice is yours. One apparent concern is that a single file introduces a single point of failure. However, we take measures in the software to make sure outages of the file do not interrupt 24*7 operations. Put simply, the archive management service in each computer acquires a copy of its configuration on first use, and keeps a copy of it locally thereafter. Consequently, outages of the shared file are tolerated. In addition, each service periodically checks for changes to the configuration and automatically reflects these changes locally
In summary, implementation planning involves:
Try to use all installation defaults; it is always simpler. Especially where port numbers are concerned.
Decide all the computers where extractors will run.
Decide all the computers where accumulators will run.
Decide which computer is to host the file for the configuration data. The specific installation sections for each platform cover the choice between sharing or establishing the file.
Install on the computer that hosts this file first. Instruct the installer to establish the file for first use.
Determine the hostname of the computer hosting the file. You need this as a reference point when doing subsequent installs. See below for information on determining hostname in various systems.
for z/OS (USS), use the nslookup command to show the “name” (hostname) of the computer
for Unix, use the nslookup command to show the “name” (hostname) of the computer
for Windows, go to the Control panel, select System and then select the Computer name tab. This will show the full computer name (hostname).
In all subsequent installs on other computers, make sure you use the shared file option, do not establish additional files.
Make sure ports are enabled where firewalls are used. Where Data Archiving for Adabas is used across multiple computers, there may be the need to adjust firewall settings. The User Interface doesn’t usually need any adjustments because communications by the User Interface are outbound. However, the runtime receives communications so it must be enabled as follows:
The following program in the Data Archiving for Adabas directory structure must be added as an exception:
adrdrv adrdrv2
The following programs in the Adabas System Coordinator directory structure must be added as exceptions:
corlc cord