Error when executing "rake db:migrate" with frozen Rails 2.1.0 on Media Temple

I’m hoping someone might have some insight into this, because I don’t
quite know how to figure this one out. I’m running the command with
Capistrano’s deploy:migrate to my (mt) Media Temple (gs) Grid Server,
but I get the same error (from “rake aborted!” and on below) when I
run “rake db:migrate” directly on the server in the application’s root
directory.

  • executing `deploy:migrate’
  • executing “ls -x /home/0000/containers/rails/babylon/releases”
    servers: [“domain.com”]
    [domain.com] executing command
    command finished
  • executing “cd /home/0000/containers/rails/babylon/releases/
    20080608045457; PATH=$PATH:/home/0000/data/rubygems/bin:/home/0000/
    data/rubygems/gems/bin RUBYLIB=/home/0000/data/rubygems/local/lib/
    site_ruby/1.8 GEM_HOME=/home/0000/data/rubygems/gems rake
    RAILS_ENV=production db:migrate”
    servers: [“domain.com”]
    [domain.com] executing command
    *** [err :: domain.com] rake aborted!
    *** [err :: domain.com] undefined method `collect’ for nil:NilClass
    *** [err :: domain.com]
    *** [err :: domain.com] (See full trace by running task with --trace)
    ** [out :: domain.com] (in /home/0000/containers/rails/babylon/
    releases/20080608045457)
    command finished
    command “cd /home/0000/containers/rails/babylon/releases/
    20080608045457; PATH=$PATH:/home/0000/data/rubygems/bin:/home/0000/
    data/rubygems/gems/bin RUBYLIB=/home/0000/data/rubygems/local/lib/
    site_ruby/1.8 GEM_HOME=/home/0000/data/rubygems/gems rake
    RAILS_ENV=production db:migrate” failed on eyequeue.net

I can provide a full trace if necessary.