how does falcon handling forking as it relates to database connections?

[josh-m-sharpe]

in puma land, we usually do something like this:

pseudocode:

before_fork do
  disconnect_database_connection
end

after_fork do
  reconnect_database_connection
end

Is there a corollary in falcon?

[josh-m-sharpe]

I see this thing: https://socketry.github.io/falcon/guides/rails-integration/index#isolation-level - but wrapping all our calls to active record in a mature codebase is infeasible.

[ioquatix]

We should update that documentation, ActiveRecord::Base.connection_pool.with_connection became the default for both threads and fibers.

Since Falcon does not fork while handling application code, so you should not need any such handlers.

[josh-m-sharpe]

Thanks. Some updated documentation would be very welcome. I think we're so used to forking severs and how to deploy those, anything different than that leaves a lot to discover :)