I hate to hijack this thread, but I was thinking about this over the holiday in how we are going to maintain a local up to date copy of Galaxy ourselves.  One thought would be to standardize on a single supported platform aka an Amazon instance and make sure everything works on that single platform.  This frees you from having to support many different platforms and versions.  Have you guys given this some thought?

On Mon, Jan 5, 2015 at 12:05 PM, Nate Coraor <nate@bx.psu.edu> wrote:
Hi Michael,

I understand your frustration in this regard. We do make an effort to respond to and fix issues that people encounter when upgrading Galaxy releases. We've also made pretty significant strides in increasing the quality of our releases, including a feature freeze for at least 2 weeks prior to each release. The galaxy-dev mailing list is fairly active, as is our IRC channel (#galaxyproject on Freenode) for more real-time help.

If you're not familiar with the Galaxy News Briefs, these are the primary means of communicating the changes you should expect to see between Galaxy releases and any config changes that might be necessary between releases:


We also recommend having a Test/QA Galaxy server (or cloning your production Galaxy instance) to test an upgrade before deploying it on your production instance(s) so most problems can be caught beforehand.

Unfortunately, due to the wide range of platforms, deployment scenarios, site-specific factors, and large scope of Galaxy functionality, we can't catch all potential issues. I don't think this is outside the rate typically experienced with major releases of large open source software, but it is still frustrating when it occurs. That said, we'll do our best to assist and improve documentation when you do encounter upgrade problems.

--nate