the first server has the ip 192.168.122.138 so I changed the config in the mariadb.cnf to match. It doesn't change anything.

No, you still need wsrep_cluster_address.
wsrep_sst_receive_address is the address of the current node, while wsrep_cluster_address is where it is supposed to connect to and receive the state from. Without it, it doesn't know anything about an existing cluster, that's why WSREP stops at

WSREP: Assign initial position for certification: 0, protocol version: -1

and doesn't proceed with replication.
Please set the cluster address back as usual.

It proceeded much further now, so we are making progress
Now it says

Feb 11 16:13:47 mariadb2 mysqld: #011Read: 'rsync daemon already running.'
Feb 11 16:13:47 mariadb2 mysqld: 130211 16:13:47 [ERROR] WSREP: Process completed with error: wsrep_sst_rsync --role 'joiner' --address '192.168.122.241' --auth '' --datadir '/var/lib/mysql/' --defaults-file '/etc/mysql/my.cnf' --parent '11358': 114 (Operation already in progress)

While I'm not sure that's what happens in your case, I've seen the error before when rsync would hang running from previous attempts.
Please check if you have rsync running on each of the machines (sender and receiver), kill it if it's hung there and try again.
If it's hanging it will probably require a brutal SIGKILL, at least that's how it used to be for me.

Okay, we had one just like that not long ago. Please check this comment and the next ones in the same report:

https://mariadb.atlassian.net/browse/MDEV-4112?focusedCommentId=29637&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-29637

You might have either the common problem with Debian passwords described at stackoverflow.com (the link provided in the comment), or the Galera-specific trick with password replication. In the latter case, the reporter described how he solved the problem, please try to follow his advice.

Good luck with your Galera cluster, hope you'll enjoy it.