Struct sqlx::pool::PoolOptions

source ·
pub struct PoolOptions<DB>where
    DB: Database,{ /* private fields */ }
Expand description

Configuration options for Pool.

Callback Functions: Why Do I Need Box::pin()?

Essentially, because it’s impossible to write generic bounds that describe a closure with a higher-ranked lifetime parameter, returning a future with that same lifetime.

Ideally, you could define it like this:

async fn takes_foo_callback(f: impl for<'a> Fn(&'a mut Foo) -> impl Future<'a, Output = ()>)

However, the compiler does not allow using impl Trait in the return type of an impl Fn.

And if you try to do it like this:

async fn takes_foo_callback<F, Fut>(f: F)
where
    F: for<'a> Fn(&'a mut Foo) -> Fut,
    Fut: for<'a> Future<Output = ()> + 'a

There’s no way to tell the compiler that those two 'as should be the same lifetime.

It’s possible to make this work with a custom trait, but it’s fiddly and requires naming the type of the closure parameter.

Having the closure return BoxFuture allows us to work around this, as all the type information fits into a single generic parameter.

We still need to Box the future internally to give it a concrete type to avoid leaking a type parameter everywhere, and Box is in the prelude so it doesn’t need to be manually imported, so having the closure return Pin<Box<dyn Future> directly is the path of least resistance from the perspectives of both API designer and consumer.

Implementations§

source§

impl<DB> PoolOptions<DB>where DB: Database,

source

pub fn new() -> PoolOptions<DB>

Returns a default “sane” configuration, suitable for testing or light-duty applications.

Production applications will likely want to at least modify max_connections.

See the source of this method for the current default values.

source

pub fn max_connections(self, max: u32) -> PoolOptions<DB>

Set the maximum number of connections that this pool should maintain.

Be mindful of the connection limits for your database as well as other applications which may want to connect to the same database (or even multiple instances of the same application in high-availability deployments).

source

pub fn get_max_connections(&self) -> u32

Get the maximum number of connections that this pool should maintain

source

pub fn min_connections(self, min: u32) -> PoolOptions<DB>

Set the minimum number of connections to maintain at all times.

When the pool is built, this many connections will be automatically spun up.

If any connection is reaped by max_lifetime or idle_timeout, or explicitly closed, and it brings the connection count below this amount, a new connection will be opened to replace it.

This is only done on a best-effort basis, however. The routine that maintains this value has a deadline so it doesn’t wait forever if the database is being slow or returning errors.

This value is clamped internally to not exceed max_connections.

We’ve chosen not to assert min_connections <= max_connections anywhere because it shouldn’t break anything internally if the condition doesn’t hold, and if the application allows either value to be dynamically set then it should be checking this condition itself and returning a nicer error than a panic anyway.

source

pub fn get_min_connections(&self) -> u32

Get the minimum number of connections to maintain at all times.

source

pub fn acquire_timeout(self, timeout: Duration) -> PoolOptions<DB>

Set the maximum amount of time to spend waiting for a connection in Pool::acquire().

Caps the total amount of time Pool::acquire() can spend waiting across multiple phases:

  • First, it may need to wait for a permit from the semaphore, which grants it the privilege of opening a connection or popping one from the idle queue.
  • If an existing idle connection is acquired, by default it will be checked for liveness and integrity before being returned, which may require executing a command on the connection. This can be disabled with test_before_acquire(false).
  • If a new connection needs to be opened, that will obviously require I/O, handshaking, and initialization commands.
source

pub fn get_acquire_timeout(&self) -> Duration

Get the maximum amount of time to spend waiting for a connection in Pool::acquire().

source

pub fn max_lifetime( self, lifetime: impl Into<Option<Duration>> ) -> PoolOptions<DB>

Set the maximum lifetime of individual connections.

Any connection with a lifetime greater than this will be closed.

When set to None, all connections live until either reaped by idle_timeout or explicitly disconnected.

Infinite connections are not recommended due to the unfortunate reality of memory/resource leaks on the database-side. It is better to retire connections periodically (even if only once daily) to allow the database the opportunity to clean up data structures (parse trees, query metadata caches, thread-local storage, etc.) that are associated with a session.

source

pub fn get_max_lifetime(&self) -> Option<Duration>

Get the maximum lifetime of individual connections.

source

pub fn idle_timeout( self, timeout: impl Into<Option<Duration>> ) -> PoolOptions<DB>

Set a maximum idle duration for individual connections.

Any connection that remains in the idle queue longer than this will be closed.

For usage-based database server billing, this can be a cost saver.

source

pub fn get_idle_timeout(&self) -> Option<Duration>

Get the maximum idle duration for individual connections.

source

pub fn test_before_acquire(self, test: bool) -> PoolOptions<DB>

If true, the health of a connection will be verified by a call to Connection::ping before returning the connection.

Defaults to true.

source

pub fn get_test_before_acquire(&self) -> bool

Get’s whether test_before_acquire is currently set.

source

pub fn after_connect<F>(self, callback: F) -> PoolOptions<DB>where F: for<'c> Fn(&'c mut <DB as Database>::Connection, PoolConnectionMetadata) -> Pin<Box<dyn Future<Output = Result<(), Error>> + Send + 'c, Global>> + 'static + Send + Sync,

Perform an asynchronous action after connecting to the database.

If the operation returns with an error then the error is logged, the connection is closed and a new one is opened in its place and the callback is invoked again.

This occurs in a backoff loop to avoid high CPU usage and spamming logs during a transient error condition.

Note that this may be called for internally opened connections, such as when maintaining min_connections, that are then immediately returned to the pool without invoking after_release.

Example: Additional Parameters

This callback may be used to set additional configuration parameters that are not exposed by the database’s ConnectOptions.

This example is written for PostgreSQL but can likely be adapted to other databases.

use sqlx::Executor;
use sqlx::postgres::PgPoolOptions;

let pool = PgPoolOptions::new()
    .after_connect(|conn, _meta| Box::pin(async move {
        // When directly invoking `Executor` methods,
        // it is possible to execute multiple statements with one call.
        conn.execute("SET application_name = 'your_app'; SET search_path = 'my_schema';")
            .await?;

        Ok(())
    }))
    .connect("postgres:// …").await?;

For a discussion on why Box::pin() is required, see the type-level docs.

source

pub fn before_acquire<F>(self, callback: F) -> PoolOptions<DB>where F: for<'c> Fn(&'c mut <DB as Database>::Connection, PoolConnectionMetadata) -> Pin<Box<dyn Future<Output = Result<bool, Error>> + Send + 'c, Global>> + 'static + Send + Sync,

Perform an asynchronous action on a previously idle connection before giving it out.

Alongside the connection, the closure gets PoolConnectionMetadata which contains potentially useful information such as the connection’s age and the duration it was idle.

If the operation returns Ok(true), the connection is returned to the task that called Pool::acquire.

If the operation returns Ok(false) or an error, the error is logged (if applicable) and then the connection is closed and Pool::acquire tries again with another idle connection. If it runs out of idle connections, it opens a new connection instead.

This is not invoked for new connections. Use after_connect for those.

Example: Custom test_before_acquire Logic

If you only want to ping connections if they’ve been idle a certain amount of time, you can implement your own logic here:

This example is written for Postgres but should be trivially adaptable to other databases.

use sqlx::{Connection, Executor};
use sqlx::postgres::PgPoolOptions;

let pool = PgPoolOptions::new()
    .test_before_acquire(false)
    .before_acquire(|conn, meta| Box::pin(async move {
        // One minute
        if meta.idle_for.as_secs() > 60 {
            conn.ping().await?;
        }

        Ok(true)
    }))
    .connect("postgres:// …").await?;

For a discussion on why Box::pin() is required, see the type-level docs.

source

pub fn after_release<F>(self, callback: F) -> PoolOptions<DB>where F: for<'c> Fn(&'c mut <DB as Database>::Connection, PoolConnectionMetadata) -> Pin<Box<dyn Future<Output = Result<bool, Error>> + Send + 'c, Global>> + 'static + Send + Sync,

Perform an asynchronous action on a connection before it is returned to the pool.

Alongside the connection, the closure gets PoolConnectionMetadata which contains potentially useful information such as the connection’s age.

If the operation returns Ok(true), the connection is returned to the pool’s idle queue. If the operation returns Ok(false) or an error, the error is logged (if applicable) and the connection is closed, allowing a task waiting on Pool::acquire to open a new one in its place.

Example (Postgres): Close Memory-Hungry Connections

Instead of relying on max_lifetime to close connections, we can monitor their memory usage directly and close any that have allocated too much.

Note that this is purely an example showcasing a possible use for this callback and may be flawed as it has not been tested.

This example queries pg_backend_memory_contexts which is only allowed for superusers.

use sqlx::{Connection, Executor};
use sqlx::postgres::PgPoolOptions;

let pool = PgPoolOptions::new()
    // Let connections live as long as they want.
    .max_lifetime(None)
    .after_release(|conn, meta| Box::pin(async move {
        // Only check connections older than 6 hours.
        if meta.age.as_secs() < 6 * 60 * 60 {
            return Ok(true);
        }

        let total_memory_usage: i64 = sqlx::query_scalar(
            "select sum(used_bytes) from pg_backend_memory_contexts"
        )
        .fetch_one(conn)
        .await?;

        // Close the connection if the backend memory usage exceeds 256 MiB.
        Ok(total_memory_usage <= (2 << 28))
    }))
    .connect("postgres:// …").await?;
source

pub async fn connect( self, url: &str ) -> impl Future<Output = Result<Pool<DB>, Error>>

Create a new pool from this PoolOptions and immediately open at least one connection.

This ensures the configuration is correct.

The total number of connections opened is max(1, min_connections).

Refer to the relevant ConnectOptions impl for your database for the expected URL format:

  • Postgres: [PgConnectOptions][crate::postgres::PgConnectOptions]
  • MySQL: [MySqlConnectOptions][crate::mysql::MySqlConnectOptions]
  • SQLite: [SqliteConnectOptions][crate::sqlite::SqliteConnectOptions]
  • MSSQL: [MssqlConnectOptions][crate::mssql::MssqlConnectOptions]
source

pub async fn connect_with( self, options: <<DB as Database>::Connection as Connection>::Options ) -> impl Future<Output = Result<Pool<DB>, Error>>

Create a new pool from this PoolOptions and immediately open at least one connection.

This ensures the configuration is correct.

The total number of connections opened is max(1, min_connections).

source

pub fn connect_lazy(self, url: &str) -> Result<Pool<DB>, Error>

Create a new pool from this PoolOptions, but don’t open any connections right now.

If min_connections is set, a background task will be spawned to optimistically establish that many connections for the pool.

Refer to the relevant ConnectOptions impl for your database for the expected URL format:

  • Postgres: [PgConnectOptions][crate::postgres::PgConnectOptions]
  • MySQL: [MySqlConnectOptions][crate::mysql::MySqlConnectOptions]
  • SQLite: [SqliteConnectOptions][crate::sqlite::SqliteConnectOptions]
  • MSSQL: [MssqlConnectOptions][crate::mssql::MssqlConnectOptions]
source

pub fn connect_lazy_with( self, options: <<DB as Database>::Connection as Connection>::Options ) -> Pool<DB>

Create a new pool from this PoolOptions, but don’t open any connections right now.

If min_connections is set, a background task will be spawned to optimistically establish that many connections for the pool.

Trait Implementations§

source§

impl<DB> Clone for PoolOptions<DB>where DB: Database,

source§

fn clone(&self) -> PoolOptions<DB>

Returns a copy of the value. Read more
1.0.0 · source§

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source. Read more
source§

impl<DB> Debug for PoolOptions<DB>where DB: Database,

source§

fn fmt(&self, f: &mut Formatter<'_>) -> Result<(), Error>

Formats the value using the given formatter. Read more
source§

impl<DB> Default for PoolOptions<DB>where DB: Database,

source§

fn default() -> PoolOptions<DB>

Returns the “default value” for a type. Read more

Auto Trait Implementations§

§

impl<DB> !RefUnwindSafe for PoolOptions<DB>

§

impl<DB> Send for PoolOptions<DB>

§

impl<DB> Sync for PoolOptions<DB>

§

impl<DB> Unpin for PoolOptions<DB>

§

impl<DB> !UnwindSafe for PoolOptions<DB>

Blanket Implementations§

source§

impl<T> Any for Twhere T: 'static + ?Sized,

source§

fn type_id(&self) -> TypeId

Gets the TypeId of self. Read more
source§

impl<T> Borrow<T> for Twhere T: ?Sized,

source§

fn borrow(&self) -> &T

Immutably borrows from an owned value. Read more
source§

impl<T> BorrowMut<T> for Twhere T: ?Sized,

source§

fn borrow_mut(&mut self) -> &mut T

Mutably borrows from an owned value. Read more
source§

impl<T> From<T> for T

source§

fn from(t: T) -> T

Returns the argument unchanged.

source§

impl<T> Instrument for T

source§

fn instrument(self, span: Span) -> Instrumented<Self>

Instruments this type with the provided Span, returning an Instrumented wrapper. Read more
source§

fn in_current_span(self) -> Instrumented<Self>

Instruments this type with the current Span, returning an Instrumented wrapper. Read more
source§

impl<T, U> Into<U> for Twhere U: From<T>,

source§

fn into(self) -> U

Calls U::from(self).

That is, this conversion is whatever the implementation of From<T> for U chooses to do.

source§

impl<T> Same<T> for T

§

type Output = T

Should always be Self
source§

impl<T> ToOwned for Twhere T: Clone,

§

type Owned = T

The resulting type after obtaining ownership.
source§

fn to_owned(&self) -> T

Creates owned data from borrowed data, usually by cloning. Read more
source§

fn clone_into(&self, target: &mut T)

Uses borrowed data to replace owned data, usually by cloning. Read more
source§

impl<T, U> TryFrom<U> for Twhere U: Into<T>,

§

type Error = Infallible

The type returned in the event of a conversion error.
source§

fn try_from(value: U) -> Result<T, <T as TryFrom<U>>::Error>

Performs the conversion.
source§

impl<T, U> TryInto<U> for Twhere U: TryFrom<T>,

§

type Error = <U as TryFrom<T>>::Error

The type returned in the event of a conversion error.
source§

fn try_into(self) -> Result<U, <U as TryFrom<T>>::Error>

Performs the conversion.
§

impl<V, T> VZip<V> for Twhere V: MultiLane<T>,

§

fn vzip(self) -> V

source§

impl<T> WithSubscriber for T

source§

fn with_subscriber<S>(self, subscriber: S) -> WithDispatch<Self>where S: Into<Dispatch>,

Attaches the provided Subscriber to this type, returning a WithDispatch wrapper. Read more
source§

fn with_current_subscriber(self) -> WithDispatch<Self>

Attaches the current default Subscriber to this type, returning a WithDispatch wrapper. Read more