Appendix A. Upgrading From Red Hat Enterprise Linux 6 High Availability Add-On

This appendix provides an overview of upgrading Red Hat Enterprise Linux High Availability Add-On from release 6 to release 7.

A.1. Overview of Differences Between Releases

The Red Hat Enterprise Linux 7 High Availability Add-On introduces a new suite of technologies that underlie high-availability systems. These technologies are based on Pacemaker and Corosync and they replace the CMAN and RGManager technologies from previous releases of the High Availability Add-On. Below are some of the differences between the two releases. For a more comprehensive look at the differences between releases, see the appendix titled "Cluster Creation with rgmanager and with Pacemaker" from the Red Hat Enterprise Linux High Availability Add-On Reference.
  • Configuration Files — Previously, cluster configuration was found in the /etc/cluster/cluster.conf file, while cluster configuration in release 7 is in /etc/corosync/corosync.conf for membership and quorum configuration and /var/lib/pacemaker/cib/cib.xml for cluster node and resource configuration.
  • Executable Files — Previously, cluster commands were in ccs by means of a command line, luci for graphical configuration. In Red Hat Enterprise Linux 7 High Availability Add-On, configuration is done by means of pcs at the command line and the pcsd Web UI configuration at the desktop.
  • Starting the Service — Previously, all services including those in High Availability Add-On were performed using the service command to start services and the chkconfig command to configure services to start upon system boot. This had to be configured separately for all cluster services (rgmanager, cman, and ricci. For example:
    service rgmanager start
    chkconfig rgmanager on
    
    For Red Hat Enterprise Linux 7 High Availability Add-On, the systemctl controls both manual startup and automated boot-time startup, and all cluster services are grouped in the pcsd.service. For example:
    systemctl start pcsd.service 
    systemctl enable pcsd.service 
    pcs cluster start -all
    
  • User Access — Previously, the root user or a user with proper permissions can access the luci configuration interface. All access requires the ricci password for the node.
    In Red Hat Enterprise Linux 7 High Availability Add-On, the pcsd Web UI requires that you authenticate as user hacluster, which is the common system user. The root user can set the password for hacluster.
  • Creating Clusters, Nodes and Resources — Previously, creation of nodes were performed with the ccs by means of a command line or with luci graphical interface. Creation of a cluster and adding nodes is a separate process. For example, to create a cluster and add a node by means of the command line, perform the following:
    ccs -h node1.example.com --createcluster examplecluster
    ccs -h node1.example.com --addnode node2.example.com
    
    In Red Hat Enterprise Linux 7 High Availability Add-On, adding of clusters, nodes, and resources are done by means of pcs at the command line, or the pcsd Web UI. For example, to create a cluster by means of the command line, perform the following:
    pcs cluster setup examplecluster node1 node2 ...
    
  • Cluster removal — Previously, administrators removed a cluster by deleting nodes manually from the luci interface or deleting the cluster.conf file from each node
    In Red Hat Enterprise Linux 7 High Availability Add-On, administrators can remove a cluster by issuing the pcs cluster destroy command.