Unable to create cluster using luci. Found logs are filling with '/usr/libexec/ricci/ricci-worker -f /var/lib/ricci/queue/XXXX' messages.

Resolution

This issue has been resolved for RHEL 6 in the following errata: RHBA-2012-0898
This issue has been resolved for RHEL 6.z in the following errata: RHBA-2011:1510
This issue has been resolved for RHEL 5 in the following errata: RHSA-2012-0151
This issue has been resolved for RHEL 5.z in the following errata:RHBA-2011:1421

Workaround

When upgrading Luci is not possible due to local policies, a *workaround * can be used. to install or update all the packages for the cluster/shared storage channels so that the nodes have all the required packages installed or updated..

Install the required packages for all the cluster nodes that will be in the new cluster before using luci to "Create a New Cluster". If the option "Enable Shared Storage Support" is not going to be enabled/checked when creating a new cluster is created then that yum group for shared storage can be omitted. In RHEL 5 the yum group is called: "Cluster Storage" and in RHEL 6 the yum group is called: "Resilient Storage".

Install all the "Clustering" packages and the "Cluster Storage" packages for RHEL 5 so that ricci will not attempt to install the packages:

$ yum -y groupinstall "Clustering" "Cluster Storage"

Install all the "High Availability", "High Availability Management" , and "Resilient Storage" packages for RHEL 6 so that ricci will not attempt to install the packages. The package luci can be removed after install if that package is not needed:

$ yum -y groupinstall "High Availability" "High Availability Management" "Resilient Storage" 

Then create the new cluster with luci after entering all the required information after clicking "Create a New Cluster" button. If creating the cluster appears to hang on the "install" phase of the cluster install then kill the yum processes(view the /var/log/messages to see if it is hung). It is possible that there will be 3 processes that will need to be killed. Once the first process is killed then another yum process is created, and after second yum process is killed then a 3rd yum process is started which will need to be killed. Here is example of the process of killing the 3 yum processes:

$ ps aux | grep yum | grep -v grep
root 19343  4.4 10.3 229908 52944 ? S 11:27 0:06 /usr/bin/python /usr/bin/yum -y list all
$ kill 19343
$ ps aux | grep yum | grep -v grep
root 19389 43.5 10.3 229904 52940 ? S 11:30 0:06 /usr/bin/python /usr/bin/yum -y list all
$ kill 19389
$ ps aux | grep yum | grep -v grep
root 19401 93.7 10.3 229488 52512 ? R 11:31 0:03 /usr/bin/python /usr/bin/yum -y list all
$ logger "Killing YUM"; kill 19401

The install should finish after that and start the cluster services if all the packages required are already installed. If this procedure fails to create the new cluster, then try using the same procedure without enabling/unchecking the "Enable Shared Storage Support" option.