db:migrate:reset doesn't work with environment variables - ruby-on-rails

Question

The way I set up environment variables is like it was outlined here.
In other words I created a config/app_env_vars.rb file and put:
unless Rails.env.production?
ENV['DB_PASSWORD'] = 'password'
ENV['DB_USERNAME'] = 'username'
end
puts 'ECHO app_env_vars.rb'
In config/environment.rb I put:
app_env_vars = File.join(Rails.root, 'config', 'app_env_vars.rb')
load(app_env_vars) if File.exists?(app_env_vars)
puts "ECHO environment.rb"
right before Rails.application.initialize!
In database.yml:
development:
adapter: postgresql
encoding: unicode
database: my_app_development
host: localhost
pool: 5
password: <%= ENV['DB_PASSWORD'] %>
test:
adapter: postgresql
encoding: unicode
database: my_app_test
host: localhost
pool: 5
password: <%= ENV['DB_PASSWORD'] %>
When I run tasks such as bundle exec rake db:migrate and bundle exec rake db:reset it works fine but when I run bundle exec rake db:migrate:reset it will fail displaying:
fe_sendauth: no password supplied
.
.
.
Couldn't drop my_app_development
fe_sendauth: no password supplied
.
.
.
Couldn't drop my_app_test
fe_sendauth: no password supplied
ECHO app_env_vars.rb
ECHO environment.rb
Usually when I run bundle exec rake db:migrate (and db:drop) it will first display the ECHO parts and then do the task and out put the log. But in the case of the db:migrate:reset it won't initialize the environment variables until after the task.
When in database.yml I put the actual passwords rather than the env variables then the
db:migrate:reset works with no problem.
Here is rake db:migrate:reset with --trace option
** Invoke db:migrate:reset (first_time)
** Invoke db:drop (first_time)
** Invoke db:load_config (first_time)
** Execute db:load_config
** Execute db:drop
fe_sendauth: no password supplied
.
. # gem trace stuff...
.
Couldn't drop my_app_development
fe_sendauth: no password supplied
.
. # gem trace stuff...
.
Couldn't drop my_app_test
** Invoke db:create (first_time)
** Invoke db:load_config
** Execute db:create
fe_sendauth: no password supplied
.
. # gem trace stuff...
.
Couldn't create database for {"adapter"=>"postgresql", "encoding"=>"unicode", "database"=>"my_app_development", "host"=>"localhost", "pool"=>5, "password"=>nil}
fe_sendauth: no password supplied
.
. # gem trace stuff...
.
Couldn't create database for {"adapter"=>"postgresql", "encoding"=>"unicode", "database"=>"my_app_test", "host"=>"localhost", "pool"=>5, "password"=>nil}
** Invoke db:migrate (first_time)
** Invoke environment (first_time)
** Execute environment
ECHO app_env_vars.rb
ECHO environment.rb
** Invoke db:load_config
** Execute db:migrate
** Invoke db:_dump (first_time)
** Execute db:_dump
** Invoke db:schema:dump (first_time)
** Invoke environment
** Invoke db:load_config
** Execute db:schema:dump
** Execute db:migrate:reset
Here is the --trace for rake db:reset
** Invoke db:reset (first_time)
** Invoke environment (first_time)
** Execute environment
ECHO app_env_vars.rb
ECHO environment.rb
** Invoke db:load_config (first_time)
** Execute db:load_config
** Execute db:reset
** Invoke db:drop (first_time)
** Invoke db:load_config
** Execute db:drop
** Invoke db:setup (first_time)
** Invoke db:schema:load_if_ruby (first_time)
** Invoke db:create (first_time)
** Invoke db:load_config
** Execute db:create
** Invoke environment
** Execute db:schema:load_if_ruby
** Invoke db:schema:load (first_time)
** Invoke environment
** Invoke db:load_config
** Execute db:schema:load
-- enable_extension("plpgsql")
-> 0.0772s
-- create_table("users", {:force=>true})
-> 0.0795s
-- add_index("users", ["email"], {:name=>"index_users_on_email", :unique=>true, :using=>:btree})
-> 0.0406s
-- initialize_schema_migrations_table()
-> 0.0829s
** Invoke db:structure:load_if_sql (first_time)
** Invoke db:create
** Invoke environment
** Execute db:structure:load_if_sql
** Invoke db:seed (first_time)
** Execute db:seed
** Invoke db:abort_if_pending_migrations (first_time)
** Invoke environment
** Execute db:abort_if_pending_migrations
** Execute db:setup
In the case of db:reset the ** Execute environment part is done before any executions, while with db:migrate:reset the ** Execute environment is done after a few executions

Answer 1

The reason why this setup (setting env variables in a file then load them trough environment.rb) doesn't work is because active record can be used outside rails therefore now every task runs with initialized rails environment.
A better solution is to use gems like figaro or dot-env which will load with every rake task. How exactly is that done with those gems I am not sure but if I get into it and find out I'll post here.