# Upgrading to v2.6

For Oban OSS users the v2.6 upgrade is a drop in replacement—there isn't
anything to do! However, Web+Pro users will need to make some changes to unlock
the goodness of engines.

## Bump Your Deps

Update Oban, Web, and Pro to the latest versions:

```elixir
[
  {:oban, "~> 2.6"},
  {:oban_web, "~> 2.6", repo: "oban"},
  {:oban_pro, "~> 0.7", repo: "oban"}
  ...
]
```

Be sure to specify **both `:oban_web` and `:oban_pro`** if you use them both.
There aren't any dependencies between Web and Pro now. That means you're free to
use Pro for workers and only include Web for Phoenix servers, etc.

## Switch to the SmartEngine

The `SmartEngine` uses centralized records to track and exchange state globally,
enabling features such as global concurrency.

First, create a migration to add the new `oban_producers` table:

```bash
$ mix ecto.gen.migration add_oban_producers
```

Within the migration module:

```elixir
use Ecto.Migration

defdelegate change, to: Oban.Pro.Migrations.Producers
```

If you have multiple Oban instances or use prefixes, you can specify the prefix
and create multiple tables in one migration:

```elixir
use Ecto.Migration

def change do
  Oban.Pro.Migrations.Producers.change()
  Oban.Pro.Migrations.Producers.change(prefix: "special")
  Oban.Pro.Migrations.Producers.change(prefix: "private")
end
```

Next, update your config to use the `SmartEngine`:

```elixir
config :my_app, Oban,
  engine: Oban.Pro.Queue.SmartEngine,
  ...
```

If you have multiple Oban instances you need to configure each one to use the
`SmartEngine`, otherwise they'll default to the `Basic` engine.

## Start Gossiping

Oban Pro no longer writes heartbeat records to `oban_beats`. Instead, any Oban
instance that runs queues must use the `Gossip` plugin to broadcast status via
PubSub.

To start, include the `Gossip` plugin in your Oban config:

```elixir
config :my_app, Oban,
  plugins: [
    Oban.Plugins.Gossip
    ...
```

With the default configuration the plugin will broadcast every 1 second. If that
is too frequent you can configure the interval:

```elixir
  plugins: [
    {Oban.Plugins.Gossip, interval: :timer.seconds(5)}
    ...
```

## Remove the Workflow Manager

Due to an improvement in how configuration is passed to workers the
`WorkflowManager` plugin is no longer needed. You can remove it from your list
of plugins:

```diff
plugins: [
  Oban.Pro.Plugins.Lifeline,
- Oban.Pro.Plugins.WorkflowManager,
  ...
```

## Remove Extra Lifeline Options

The `Lifeline` plugin is simplified and doesn't accept as many configuration
options. If you previously configured the `record_interval` or `delete_interval`
you can remove them:

```diff
plugins: [{
  Oban.Pro.Plugins.Lifeline,
- delete_interval: :timer.minutes(10),
- record_interval: :timer.seconds(10),
  rescue_interval: :timer.minutes(5)
}]
```

## Drop the Beats Table

Once you've rolled out the switch to producer records, the smart engine and the
gossip plugin you are free to remove the `oban_beats` table at your discretion
(preferably in a follow up release, to prevent errors):

```bash
$ mix ecto.gen.migration drop_oban_beats
```

Within the generated migration module:

```elixir
use Ecto.Migration

def up do
  drop_if_exists table("oban_beats")
  drop_if_exists table("oban_beats", prefix: "private") # If you have any prefixes:
end

def down do
  # No going back!
end
```