8.2.1 Repository Groups
Configuration of repository synchronization requires repository groups to be created. A repository group is a list of all the PCs that use a non-IBM i repository. Thus a repository group is synonymous with the users of a non-IBM i repository.
The non-IBM i repositories can be either on a stand-alone workstation or on a server in a network configuration.
Because the repository groups reflect the actual configuration of the repositories, each PC can only be in one repository group.
How you set up repository groups depends on how the PCs and the repositories are connected to the IBM i repository:
- In the simplest case, you create a repository group for a single PC which has a repository (so called stand-alone workstation configuration).
- In a configuration with a network server containing a repository and several client PCs you create one repository group to include both the server and the PCs.
In a repository group, one PC must be identified as the repository gateway. This PC uses Host Monitor to receive the propagations and store them in the repository.
In a stand-alone configuration, the single PC is the gateway as there is only one PC in the repository group. In a network configuration, the PC chosen as the gateway must have the host monitor installed, but it does not have to be the PC with the repository. In this configuration, the gateway is often the server.
If the PC identified as the repository gateway is deleted without another PC being identified as the new repository gateway, then the first PC in the repository group is considered to be the repository gateway for the repository group.
A PC must be in a repository group in order for its repository to receive any propagations.
The Host Monitor must be running on the repository gateway for changes to the current partition to be received. Propagations are queued on the IBM i until the Host Monitor is started in the appropriate partition.
A fundamental part of repository synchronization is the IBM i. Without an IBM i, no propagations can occur. If Visual LANSA is configured in standalone mode - disconnected from any IBM i - it is also disconnected from all other non-IBM i repositories.
In order to configure repository groups you need to answer at least the following questions:
- What non-IBM i repositories are there that need to take part in repository synchronization?
This gives you the repository groups, one for each repository.
- What PC will be the gateway for each repository?
The gateway is the PC that will be sent the propagations for a particular repository. This is the first member of the repository group.
- What other PCs use each repository?
This gives you the other members of each repository group.
Armed with this information you can now create your repository groups as described in 8.3.1 Add Repository Groups.
Provided you only need to add repository groups and not create PC definitions, a refresh of PC definitions is not required. The next Check In made from a PC in a repository group will be propagated to the other repositories.
A further refinement can be added to restrict which changes go to which repositories. This is referred to as a work group. When no work groups are defined, all propagations go to all repositories.