On Fri, 9 Oct 2009, Dan Ritter wrote: > On Fri, Oct 09, 2009 at 02:36:02PM -0400, David Allan wrote: >> I've heard more than a few folks say that Ruby applications present a >> unique set of challenges to deploy, update and scale in an elegant way. >> I stopped doing serious admin work before Ruby became widely used, so I >> can't speak to managing it, but as a developer I can see the appeal of >> writing in it. Would anyone who's deployed Ruby applications in >> production be willing to talk about what makes a Ruby application easy or >> hard to manage and what the gotchas are when Ruby goes into production? >> Would anybody else find that an interesting topic? > > My company's new release is a Ruby application; I can speak to the > difficulties of dealing with it, but I don't think it's worth a full > presentation. > > It boils down to this: Ruby modules, called gems, are difficult to > maintain in a consistent state. There are dependencies between: > > - exact flavor and version of ruby > - each installed gem > - mod_ruby, aka mod_passenger > > Each of these items is under constant external development, and the naive > install procedures don't guarantee a consistent environment, just an > up-to-date one. This is great for living on the edge; not so > wonderful for producing a farm of servers that all act the same > way at the same time. Linux distributors don't react quickly > enough, so that role pretty much has to be assumed by us. > > Handling this properly is a Small Matter of having consistent procedures > and maintaining your own repositories. A good relationship with your > developers is, of course, essential. They need to be able to trust > IT to turn around requests quickly, and to trust operations to follow > documented procedures. But then, what developers don't need those things?
That's all consistent with what I know and what I would expect, and I'm sure it's not probably all that tricky, but where I was going with this is that I am somewhat unfamiliar--and I suspect others are as well--with what Ruby, Rails, gems, mongrel, etc, are, and are also unfamiliar with best practices for integrating everything into that farm of servers that all behave alike that we all want, what parameters are likely to indicate that the health of the system is not perfect, what the first performance bottlenecks are likely to be, etc. If you or anybody else would be up for talking about that stuff, I at least would be interested in hearing the details. Dave _______________________________________________ bblisa mailing list [email protected] http://www.bblisa.org/mailman/listinfo/bblisa
