I am told that the structure of the channels for RHC 1.2.3 is that calamari-server and -clients, and their dependencies, exist in the same place as the calamari-minions dependencies. The former is "salt/zmq/python libs", and the latter is "salt/zmq/Diamond", so there *is* a lot of overlap; however, it's only our current configuration limitations that mean the minions repo should even be the same distro/arch as the calamari server code, and certainly logically those are very different repos. Ideally we'd have one repo for calamari-server/clients and its dependencies, and a separate one 'calamari-minions' for things required for the minion machines (Diamond and salt et. al.). This is what ice-tools constructs, and will construct for non-RHEL distros.
QE: This bug was just to re-arrange the package mappings for 1.3 (versus 1.2). In order to move this bug to VERIFIED, you can simply verify that Calamari functions properly when only the "mon" or "osd" repos are enabled. To be specific, to test this bug: 1) Your monitor nodes have the "ceph-mon" repository enabled (and not "ceph-osd"), 2) Your OSD nodes have the "ceph-osd" repository enabled (and not "ceph-mon"), And you should see that: 3) "ceph-deploy calamari connect <mon-node>" works, 4) "ceph-deploy calamari connect <osd-node>" works, 5) The "salt-minion" package gets installed on your monitor and OSD nodes, 6) The "diamond" package gets installed on your monitor and OSD nodes, 7) and the monitors and OSDs correctly report into your Calamari admin node.
Works Fine with the ISO Installation.. Moving to verified state..
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2015:1183