Re: Is this not relabelling?
I've worked with BrokenStack everyday at work since mid-2013. None of the major distros (RH, Mirantis, Canonical or Rackspace) are by default production un-usable. Upgrades are horrible. Lifecycle is horrible. Scaling is horrible. Federation is horrible. Piece of the solution fail spectacularly. You need sumologic/loggly plus ELK plus Riemann just to make sense of all the trash logs generated by BrokenStack. What I am seeing more now is companies are bringing their networks/routers to Amazon/Google/Azure and getting sub-private-cloud pricing on storage and compute while keeping the bandwidth bills under control by using their own networks. Anyways, the last 2.5 years of my life have been most likely wasted on working on BrokenStack. Red Hat, normally the bastion of sanity in the opensource bazaar, has gone off the rails a bit and has done nothing to productize brokenstack. Mirantis simply wants you to use Fuel to install a mostly working setup and then be slave to their professional services. Canonical is interesting but they rip and replace and change everything ever 9-12 months so its hard to use them over time. Rackspace is ok - their model is run the playbooks and everything gets fixed and that is nice but upgrades of major releases is a bear. After working with brokenstack for this long I'll know what the private cloud POC needs to look like in my next job. I know exactly how to make all these snake oil vendors collapse with a few simple test cases for the POC.
BrokenStack is also largely a virtualization platform which fails to consider nesting within VMs, containers or bare-metal services.
I have a new word for brokenstack - openshitack