I've got a what I thought was a relatively simple DS setup with two master DS servers doing master-master replication. In the future slaves may be added into the equation.
Initially I installed both servers the same - as standalone DS' each with it's own admin server and 'in-house' o=NetscapeRoot configuration directory.
Reading some of the Redhat docs on 'Configuration decisions' it talks about having the configuration directory in a separate directory instance - based on what I've seen from the DS setup script this implies supplying those details during the install of the real DS instances that will contain the data.
Is my understanding correct? Does this mean I should be installing an independent configuration directory on both masters and setup replication between them to provide a redundant configuration directory alongside the redundant data directories?
If so is the install procedure reasonable?:
- install fedora RPM on server 1
- Run setup script to create server 1 config directory
- Run ds_newinst.pl to create data directory on the same server pointing it to the local config directory during setup
- Repeat on server 2
- Setup replication on data masters and on config directories
Many thanks!
Are you the Quizmaster? Play BrainBattle with a friend now!
-- Fedora-directory-users mailing list Fedora-directory-users@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-directory-users
