This code was taken from ActiveRecord 2.3.14's gem class ConnectionHandler
def establish_connection(name, spec)
@connection_pools[name] = ConnectionAdapters::ConnectionPool.new(spec)
end
It seems each time ruby calls establish_connection
on the model, it's creating a new connection pool.
My question:
If I have 5 models that use establish_connection
to the same database, is Rails smart enough to pick an already existing pool rather creating a new one with the same connection credentials? Does this also happen if my 5 models are subclasses of some abstract class that uses establish_connection
? Will it always pick a connection from the @connection_pools
if it exists?
Update 1
I'm talking about a concrete example. You have 5 models with 5 different connections, each time Rails uses a model it executes establish_connection
. Looking at the code in ActiveRecord, when it executes establish_connection
it creates a new pool with connections to that specific connection. What I'm wondering is whether each time Rails calls a model's establish_connection
, does it create a new pool or take the existing one.
Example: you come to my site and see a product list. You've just hit an action that calls Product.all
, which executes establish_connection
to some database on Amazon. Then, I come to the product list, what happens? Do I grab the established connection or am I creating a new pool with that connection?
Update 2
My guess is that first time Rails loads my models it's creating pools with different connections. After, when I use some Model.method
, it just grabs the connection associated with the model and executes the method.
I'm not sure what happens when 2 models have two equal connections (not in the abstract class but in self class). Will this produce two same connection pools, or is ActiveRecord smart enough to catch this case?
See Question&Answers more detail:
os 与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…