Setting up mirroring has quite a few manual steps, although each step in itself is quite simple. The first step is to give each host a name, and make sure that the hosts can all talk to one another by these names. The second step is to enable mirroring on each server.
We start by installing Gitolite as normal on all the servers. Then, we edit each server's $HOME/.gitolite.rc file and uncomment the line defining the HOSTNAME, choosing a simple hostname as the value.
Tip
The hostname should be a simple word; it is of interest only to Gitolite and does not have any relation to the server's hostname as known by the operating system and DNS records.
Gitolite mirroring works by allowing the servers to communicate with each other using ssh keypairs named after these hostnames. This is the same mechanism that users use, except that the public key file for a server is named server-, followed by the hostname of that server. All the keys will need to be provided to all the servers...