Struct WalRecovery 

pub struct WalRecovery { /* private fields */ }

Handles WAL recovery after a crash.

Implementations

impl WalRecovery

pub fn new(dir: impl AsRef<Path>) -> Self

Creates a new recovery handler for the given WAL directory.

pub fn from_wal(wal: &WalManager) -> Self

Creates a recovery handler from a WAL manager.

pub fn read_checkpoint_metadata(&self) -> Result<Option<CheckpointMetadata>>

Reads checkpoint metadata if it exists.

Returns None if no checkpoint metadata is found.

pub fn checkpoint(&self) -> Option<CheckpointMetadata>

Returns the checkpoint metadata, if any.

This is useful for determining whether to perform a full or incremental recovery.

pub fn recover(&self) -> Result<Vec<WalRecord>>

Recovers committed records from all WAL files.

Returns only records that were part of committed transactions. If checkpoint metadata exists, only replays files from the checkpoint sequence onwards.

Errors

Returns an error if recovery fails.

pub fn recover_as<R: WalEntry>(&self) -> Result<Vec<R>>

Recovers committed records of a specific type from all WAL files.

This is the generic version of recover. Use it when recovering a WAL that stores a custom record type.

Errors

Returns an error if recovery fails.

pub fn recover_from_checkpoint( &self, checkpoint: Option<&CheckpointMetadata>, ) -> Result<Vec<WalRecord>>

Recovers committed records, starting from a specific checkpoint.

This can be used for incremental recovery when you want to skip WAL files that precede the checkpoint.

Errors

Returns an error if recovery fails.

pub fn recover_file(&self, path: impl AsRef<Path>) -> Result<Vec<WalRecord>>

Recovers committed records from a single WAL file.

Errors

Returns an error if recovery fails.

pub fn recover_file_as<R: WalEntry>( &self, path: impl AsRef<Path>, ) -> Result<Vec<R>>

Recovers committed records of a specific type from a single WAL file.

Errors

Returns an error if recovery fails.