Help with debugging sqlite's 'busy' error with Ecto3

Questions & Help Questions
, ,
8

I should’ve inspected more

Basically, the tests end up opening multiple connections anyway. For TestRepo:

{:connect,
 [
   pool_index: 1,
   repo: Ecto.Integration.PoolRepo,
   telemetry_prefix: [:ecto, :integration, :pool_repo],
   otp_app: :ecto,
   timeout: 15000,
   adapter: Sqelect,
   database: "/tmp/test_repo.db",
   pool_size: 10,
   migration_lock: nil,
   pool_size: 1
 ]}
{:connect,
 [
   pool_index: 1,
   repo: Ecto.Integration.TestRepo,
   telemetry_prefix: [:ecto, :integration, :test_repo],
   otp_app: :ecto,
   timeout: 15000,
   adapter: Sqelect,
   database: "/tmp/test_repo.db",
   migration_lock: nil,
   pool_size: 1
 ]}
{:connect,
 [
   pool_index: 6,
   repo: Ecto.Integration.PoolRepo,
   telemetry_prefix: [:ecto, :integration, :pool_repo],
   otp_app: :ecto,
   timeout: 15000,
   adapter: Sqelect,
   database: "/tmp/test_repo.db",
   pool_size: 10,
   migration_lock: nil,
   pool_size: 1
 ]}
...

For a total of 11 connections. And in my code I end up starting 11 sqlite gen servers which … is not ideal :slight_smile: