Changing from Ubuntu to CentOS can cause a lot of paths to change because the packaging is different in various places. That can cause a lot of issues. Normally I wouldn’t recommend manually changing answer files, but in this highly unusual thing I’d recommend to go over it and see which are values you’ve changed. If you didn’t, the delete the line. Since it’s a fresh server, that should still be rather safe.
Note that the installer also has a --noop flag which will show you what it would change.
In the future, it’s also helpful to get the log from /var/log/foreman-installer.
Maybe my postgres installation is the real problem.
I installed it from scratch. (It is not the version shipped with CentOS!)
The standard (postgres) path is: /var/lib/pgsql/11/data
But the foreman-installer is looking for this path: /var/lib/pgsql/data
So I created a sym link from /var/lib/pgsql/11/data to /var/lib/pgsql/data
But after the installer has run the link is vanished and a real empty directory has been created.
Maybe this line from /var/log/foreman-installer/foreman.log helps: [DEBUG 2020-04-22T09:34:41 main] Hook /usr/share/foreman-installer/hooks/boot/10-reset_foreman_db.rb returned #<Clamp::Option::Definition:0x000000018c5130 @switches=["--reset-foreman-db"], @type=:flag, @description="Drop foreman database first? You will lose all data! Unfortunately we\ncan't detect a failure at the moment so you should verify the success\nmanually. e.g. dropping can fail when DB is currently in use.", @default_value=false, @multivalued=nil, @attribute_name="reset_foreman_db"> [DEBUG 2020-04-22T09:34:41 main] Hook /usr/share/foreman-installer/hooks/boot/11-detailed_exitcodes.rb returned #<Clamp::Option::Definition:0x000000018cbfa8 @switches=["--detailed-exitcodes"], @type=:flag, @description="Provide transaction information via exit codes, see puppet-agent(8)\nfor full details.", @default_value=false, @multivalued=nil, @attribute_name="detailed_exitcodes">
[root@puppetmaster ~]# foreman-installer
/Stage[main]/Foreman::Database/Foreman::Rake[db:migrate]/Exec[foreman-rake-db:migrate]: Failed to call refresh: '/usr/sbin/foreman-rake db:migrate' returned 1 instead of one of 
/Stage[main]/Foreman::Database/Foreman::Rake[db:migrate]/Exec[foreman-rake-db:migrate]: '/usr/sbin/foreman-rake db:migrate' returned 1 instead of one of 
But if I use only the command: /usr/sbin/foreman-rake db:migrate
Hello @mcorr !
Thanks for the suggestion.
i did encounter this because i wanted to “clone” a foreman/katello instance (version 2.3/3.18), to import content from the instance we daily use, and then perform upgrade.
I didn’t find such a “satellite-clone” nor such a “instance migrate” command in foreman.