Struct BackupNode 

pub struct BackupNode<'a> { /* private fields */ }

The read-only handle on a pair’s backup node.

Built from the backup’s own gateway surface, its own deploy surface and, when the target gives the backup a file transport, the read half of that transport. See the module doc for why no method here writes.

Implementations

impl<'a> BackupNode<'a>

pub fn new( gateway: &'a dyn GatewayApi, deploy: &'a dyn DeployApi, reader: Option<&'a dyn TreeReader>, ) -> Self

A handle over the backup’s surfaces. reader is None for a backup that declares no file transport (REST only).

pub fn has_files(&self) -> bool

Whether the backup’s files can be read (it declares a transport).

pub async fn status(&self) -> Result<RedundancyStatusWire, CoreError>

The backup’s own redundancy status.

pub async fn deployment_mode(&self) -> Result<Option<String>, CoreError>

The backup’s active deployment mode, normalized (None is no mode).

pub async fn read_back( &self, expected: &Expected, ) -> Result<BTreeMap<String, Remote>, CoreError>

Each declared, comparable resource read from the backup’s REST effective view (the same read the master’s verify makes).

pub async fn read_files( &self, scope: &Scope, ) -> Option<Result<ObservedFiles, CoreError>>

The backup’s live files within scope; None when the backup declares no file transport (a REST-only backup).

pub async fn secret_names( &self, provider: &str, ) -> Result<SecretNameListing, CoreError>

The names a secret provider lists on the backup. Names only: the listing carries no value.

pub async fn sync_exhausted(&self, since_ms: i64) -> Result<bool, CoreError>

Whether the backup’s log says the gateway gave up syncing at or after since_ms (D-10, K5 d, WR-01). The lines are matched in memory and dropped; only the answer leaves.

since_ms is the epoch-ms the engine took (see wall_clock_ms) just before this run’s commit, so a line from a previous deploy or a past recovery never ends this run’s wait. The bound is applied twice: as the log query’s startTime and again on each entry’s own timestamp, which is the gateway’s clock.

Clock assumption: since_ms comes from the CLI host’s clock and the entries from the gateway’s, so they must agree to within the age of the stale line (seconds to minutes for a real previous run). A gateway clock that runs behind the host by more than the time the sync takes to give up can hide a fresh line; that only costs the early exit, and the wait ends on its budget with the one resync as before.