# RavenDB Gateway Storage

<!-- Version variant: gatewayravendb_2; default: [/nservicebus/gateway/ravendb/index.md](/nservicebus/gateway/ravendb/index.md) -->


Provides deduplication storage for the [gateway component](/nservicebus/gateway/index.md) in RavenDB.

## Usage

<!-- snippet: DefaultUsage -->

```cs
var gatewayConfiguration = new RavenGatewayDeduplicationConfiguration((builder, _) =>
{
    var documentStore = new DocumentStore
    {
        Urls = new[] { "database-server-url" },
        Database = "default-database-name"
    };

    documentStore.Initialize();

    return documentStore;
})
{
    // When running in a cluster, enable cluster wide transaction support
    EnableClusterWideTransactions = true,
};

var gatewaySettings = endpointConfiguration.Gateway(gatewayConfiguration);
```

<!-- endsnippet -->

## Cleaning up old records

After a certain amount of time, duplicates are no longer likely and deduplication data should be cleaned up. The RavenDB gateway storage component provides a built-in mechanism based on the RavenDB Expiration feature which is activated by default on the database server.

Deduplication data is kept for 7 days by default, and the cleanup job is executed every 10 minutes. To customize the expiration policy use `DeduplicationDataTimeToLive` and `FrequencyToRunDeduplicationDataCleanup`, as shown below:

<!-- snippet: CustomExpiration -->

```cs
gatewayConfiguration.DeduplicationDataTimeToLive = TimeSpan.FromDays(15);
gatewayConfiguration.FrequencyToRunDeduplicationDataCleanup = 86400;
```

<!-- endsnippet -->

In this example, deduplication data are preserved for 15 days and cleanup is run every 24 hours.

> [!NOTE]
> While it is possible to use the same deduplication database for all endpoints within a single [logical site](/nservicebus/gateway/index.md#logically-different-sites), the gateway assumes that _different_ logical sites (which are generally physically separated as well) will use separate storage infrastructure. Sending a message to multiple sites will result in messages with the same message ID delivered to each site, if those sites share a single deduplication table, the deduplication will not work correctly. In that case, separate the storage by using different database names.