Having ran a local copy of OpenShift Origin using VirtualBox I decided to have a go at running it on a Digital Ocean Droplet running CentOS 6.5. This should be easy as there is an installer which claims;
You’re one shell command away from deploying your own Platform as a Service.
Well, sort of, after a few failed installs I finally managed it using the following (this assumes you are running a clean and minimal CentOS 6.5 installation)
yum update -y
yum install -y http://www.mirrorservice.org/sites/dl.fedoraproject.org/pub/epel/6/i386/epel-release-6-8.noarch.rpm
cat > /etc/yum.repos.d/openshift-origin-deps.repo <<”EOF”
name=OpenShift Origin Dependencies — EL6
yum install -y ruby193-ruby vim-enhanced unzip curl puppet httpd-tools augeas bind bind-utils
vim /etc/selinux/config # DigitalOcean disable SELinux so change this to enabled
reboot # Reboot to enable SELinux
scl enable ruby193 bash # Make sure that all Ruby is piped through the SCL installed Ruby 1.93 binary
sh <(curl -s https://install.openshift.com/) # See notes below before answering any questions !!!
So, the problem I was having with is that by default oo_install assumes you are running Fedora 19 and populates repos_base with http://mirror.openshift.com/pub/origin-server/release/3/fedora-19/ what the install instructions fail to say is that you need to replace this with http://mirror.openshift.com/pub/origin-server/release/3/rhel-6 when prompted about the subscription configuration oo_install will use for the deployment.
You can view the output of a full oo_install run through listed in this Gist.
So, if I had run through this on a Fedora 19 Droplet, would this have been any different? In short, yes;
yum update -y
yum install -y ruby vim-enhanced unzip curl puppet httpd-tools augeas bind bind-utils
vim /etc/selinux/config # Change from disabled to enabled
sh <(curl -s https://install.openshift.com/)