无法在生产环境中访问rails控制台中的不同数据库

Raj*_*rma 6 ruby ruby-on-rails

我有database.yml,

数据库.yml

default: &default
  adapter: mysql2
  encoding: utf8
  pool: <%= ENV.fetch("RAILS_MAX_THREADS") { 5 } %>
  socket: /var/run/mysqld/mysql.sock

development:
  tp:
    <<: *default
    database: tp
    host: xxx.xxx.xxx.xxx
    username: test
    password: test
    migrations_paths: db/tp_migrate
  mi:
    <<: *default
    database: mi
    host: xxx.xxx.xxx.xxx
    username: test
    password: test
    migrations_paths: db/mi_migrate


production:
  tp:
    <<: *default
    database: tp
    host: xxx.xxx.xxx.xxx
    username: test
    password: test
    migrations_paths: db/tp_migrate
  mi:
    <<: *default
    database: mi
    host: xxx.xxx.xxx.xxx
    username: test
    password: test
    migrations_paths: db/mi_migrate
Run Code Online (Sandbox Code Playgroud)

tp db 有一个名为servant 的表。mi db 有一个名为lord 的表。我的模型为,

应用程序记录.rb

class ApplicationRecord < ActiveRecord::Base
  self.abstract_class = true
  connects_to database: { writing: :mi, reading: :mi }
end
Run Code Online (Sandbox Code Playgroud)

tp_base.rb

class TpBase < ActiveRecord::Base
  self.abstract_class = true
  connects_to database: { writing: :tp, reading: :tp }
end
Run Code Online (Sandbox Code Playgroud)

楼主.rb

class Landlord < ApplicationRecord
end
Run Code Online (Sandbox Code Playgroud)

仆人.rb

class servant < TpBase
end
Run Code Online (Sandbox Code Playgroud)

当我打开开发控制台时,我可以访问表,即仆人和房东。

开发控制台

raj@Raj:~/Desktop/testing_app(development)$ RAILS_ENV=development rails c
Running via Spring preloader in process 51558
Loading development environment (Rails 6.0.4.1)
2.6.3 :001 > Landlord.first
  Landlord Load (263.3ms)  SELECT `landlord`.* FROM `landlord` ORDER BY `landlord`.`id` ASC LIMIT 1
 => #<Landlord id: 44> 
2.6.3 :002 > Servant.first
  Servant Load (276.7ms)  SELECT `servant`.* FROM `servant` ORDER BY `servant`.`id` ASC LIMIT 1
 => #<servant id: 1> 
Run Code Online (Sandbox Code Playgroud)

当我打开生产控制台时,我只能访问仆人表,但无法访问房东表。

生产控制台

raj@Raj:~/Desktop/testing_app(development)$ RAILS_ENV=production rails c
Loading production environment (Rails 6.0.4.1)
2.6.3 :001 > Landlord.first
  Landlord Load (248.7ms)  SELECT `landlord`.* FROM `landlord` ORDER BY `landlord`.`id` ASC LIMIT 1
Traceback (most recent call last):
        1: from (irb):1
ActiveRecord::StatementInvalid (Mysql2::Error: Table 'tp.landlord' doesn't exist)
2.6.3 :002 > Servant.first
  Servant Load (265.9ms)  SELECT `servant`.* FROM `servant` ORDER BY `servant`.`id` ASC LIMIT 1
 => #<Servant id: 1> 
Run Code Online (Sandbox Code Playgroud)

但是当我在生产控制台中手动为 mi 数据库建立连接时,只能访问楼主表

手动连接

ActiveRecord::Base.built_connection(:mi).connection

raj@Raj:~/Desktop/testing_app(development)$ RAILS_ENV=production rails c
Loading production environment (Rails 6.0.4.1)
2.6.3 :001 > Landlord.first
  Landlord Load (264.9ms)  SELECT `landlord`.* FROM `landlord` ORDER BY `landlord`.`id` ASC LIMIT 1
Traceback (most recent call last):
        1: from (irb):1
ActiveRecord::StatementInvalid (Mysql2::Error: Table 'tp.landlord' doesn't exist)
2.6.3 :002 > Servant.first
  Servant Load (263.6ms)  SELECT `servant`.* FROM `servant` ORDER BY `servant`.`id` ASC LIMIT 1
 => #<Servant id: 0, name: "", phone: "", company: "", active: "\x00", has_license: "\x01", license_class: "", default_truck_id: 8, gps_color: "00FFFF", gps_device_id: nil, rate_per_hour: 0.0, rate_per_km: 0.0, rate_per_drop: 0.0, rate_per_run: 0.0, exclude_accounting: "\x00"> 
2.6.3 :003 > ActiveRecord::Base.establish_connection(:mi).connection
2.6.3 :004 > Landlord.first
  Landlord Load (243.3ms)  SELECT `landlord`.* FROM `landlord` ORDER BY `landlord`.`id` ASC LIMIT 1
 => #<Landlord id: 44>
 2.6.3 :006 > Servant.first
  Servant Load (243.7ms)  SELECT `servant`.* FROM `servant` ORDER BY `servant`.`id` ASC LIMIT 1
 => #<Servant id: 1> 
Run Code Online (Sandbox Code Playgroud)

我的问题是,为什么在生产模式下控制台默认不加载数据库?我做错了什么还是这是预期的行为,这里没有任何问题?

Chr*_*ris 6

看起来您混淆了两个概念:复制和分片。使用复制时,您需要使用主节点和只读副本的connects_to database: { writing: :tp, reading: :tp }位置。如果您没有副本,则只需指定writing:reading:writing:

查看您想要分片的代码。在这种情况下你应该使用connects_to shards: { }

例如:

class ApplicationRecord < ActiveRecord::Base
  self.abstract_class = true
  connects_to shards: { 
    default: { writing: :mi }
  }
end


class TpBase < ActiveRecord::Base
  self.abstract_class = true
  connects_to shards: { 
    shard: { writing: :tp }
  }
end
Run Code Online (Sandbox Code Playgroud)

更多信息: https://edgeguides.rubyonrails.org/active_record_multiple_databases.html#horizo ​​ntal-sharding

但不确定这是否能解决您的问题。