As I understand it from the many hundreds of deploys I've done over the last year or so that everything cycles at once. The new app is built 'offline' and then swapped in capistrano style.
If you weren't using MondoDB you would have to worry about the gap between a slug being compiled and the migrations finishing. Neil Middleton http://about.me/neilmiddleton On Wednesday, 16 February 2011 at 16:45, dblock wrote: > I am looking at building a production environment for continuous > deployment to Heroku.Traditionally we did this with two (or more) > servers and some load-balancer in front of it. We'd take one server > out, push code, bring it back in. Then we'd repeat with the second > server. Am I correct that this is not necessary with Heroku (or it > does that for me)? > > I'd like to understand what happens during deployment from the request > point-of-view. Lets say I have N dynos and am pushing new code. When > does the application get swapped? Is there any request downtime? > > We're backed by MongoDB, so we have no data migrations. The old code > works with the database and the new code works with the same database. > We can be careful about (re)saving data so that the old version > doesn't break. > > Thx > -dB. > > -- > You received this message because you are subscribed to the Google Groups > "Heroku" group. > To post to this group, send email to heroku@googlegroups.com. > To unsubscribe from this group, send email to > heroku+unsubscr...@googlegroups.com. > For more options, visit this group at > http://groups.google.com/group/heroku?hl=en. > -- You received this message because you are subscribed to the Google Groups "Heroku" group. To post to this group, send email to heroku@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
