I have checked all these items and they are indeed the way they are supposed to be.
Splunk services are running as root, so privileges should not be an issue.
The weird thing is that if I try accessing app on a standalone search-head (which is part of the cluster, by accessing the search head directly using the hosts's URL), the dashboards populate fine. Just the cluster URL (which goes through a load balancer), does not seem to want to populate the dashboards.
Also, in a SH cluster scenario, how are the Unix groups populated ? I created a group "All Unix Hosts" and added all my *nix hosts to that group, but as I connect to different nodes in the search-head, the group definition seems to disappear. It appears that this information is not replicated amongst the different search-head nodes in the cluster.