feat(table-catalog): add maintenance quarantine operations (#4201)

Co-authored-by: Henry Guo <marshawcoco@users.noreply.github.com>
This commit is contained in:
Henry Guo
2026-07-02 23:30:44 +08:00
committed by GitHub
parent d1db9a10cd
commit 31c026dd4f
8 changed files with 472 additions and 4 deletions
@@ -88,6 +88,7 @@ catalog extension.
| Manifest/data/delete reachability cleanup | Supported | Reads manifest-list and manifest Avro references, reports reachable objects, and deletes only unreferenced table objects that pass the safety window. |
| Maintenance worker run endpoint | Preview / controlled | Supports run-once execution, current-job backpressure, retry deferral, lease expiry recovery, and heartbeat updates. |
| Maintenance scheduler guardrails | Preview / controlled | Exposes disabled, paused, ready, active-job backpressure, retry deferral, quarantine boundary, recommended actions, and recent maintenance job audit timeline state for external schedulers and operators. |
| Maintenance quarantine operations | Preview / controlled | Lets operators inspect, release, retry, or abandon the current quarantined maintenance job without moving the table pointer. |
| Compaction planning | Preview / controlled | Plans partition-local and sort-order-local binpack candidates for Parquet files and does not mix data files from different partition directories or sort orders in one rewrite group. |
| Compaction commit | Preview / controlled | Can commit a safe partition-local Parquet rewrite through the catalog while preserving Iceberg data file sort order IDs in the rewritten manifest. |
| Built-in periodic scheduler | Not claimed | Operators can trigger worker runs, but continuous in-process scheduling is not claimed. |
@@ -138,6 +138,7 @@ const TABLE_CATALOG_ENDPOINTS: &[&str] = &[
"GET /{warehouse}/namespaces/{namespace}/tables/{table}/maintenance/scheduler",
"POST /{warehouse}/namespaces/{namespace}/tables/{table}/maintenance/worker/run",
"POST /{warehouse}/namespaces/{namespace}/tables/{table}/maintenance/jobs/{job}/heartbeat",
"POST /{warehouse}/namespaces/{namespace}/tables/{table}/maintenance/jobs/{job}/quarantine",
"GET /{warehouse}/namespaces/{namespace}/tables/{table}/catalog/export",
"POST /{warehouse}/namespaces/{namespace}/tables/{table}/catalog/import",
"GET /{warehouse}/namespaces/{namespace}/tables/{table}/catalog/external",
@@ -183,6 +184,7 @@ static GET_TABLE_MAINTENANCE_JOB_HANDLER: GetTableMaintenanceJobHandler = GetTab
static GET_TABLE_MAINTENANCE_SCHEDULER_HANDLER: GetTableMaintenanceSchedulerHandler = GetTableMaintenanceSchedulerHandler {};
static RUN_TABLE_MAINTENANCE_WORKER_HANDLER: RunTableMaintenanceWorkerHandler = RunTableMaintenanceWorkerHandler {};
static HEARTBEAT_TABLE_MAINTENANCE_JOB_HANDLER: HeartbeatTableMaintenanceJobHandler = HeartbeatTableMaintenanceJobHandler {};
static TABLE_MAINTENANCE_QUARANTINE_HANDLER: TableMaintenanceQuarantineHandler = TableMaintenanceQuarantineHandler {};
static EXPORT_TABLE_CATALOG_HANDLER: ExportTableCatalogHandler = ExportTableCatalogHandler {};
static IMPORT_TABLE_CATALOG_HANDLER: ImportTableCatalogHandler = ImportTableCatalogHandler {};
static EXTERNAL_CATALOG_BRIDGE_HANDLER: ExternalCatalogBridgeHandler = ExternalCatalogBridgeHandler {};
@@ -959,6 +961,11 @@ fn register_table_catalog_prefix_routes(r: &mut S3Router<AdminOperation>, prefix
format!("{prefix}/{{warehouse}}/namespaces/{{namespace}}/tables/{{table}}/maintenance/jobs/{{job}}/heartbeat").as_str(),
AdminOperation(&HEARTBEAT_TABLE_MAINTENANCE_JOB_HANDLER),
)?;
r.insert(
Method::POST,
format!("{prefix}/{{warehouse}}/namespaces/{{namespace}}/tables/{{table}}/maintenance/jobs/{{job}}/quarantine").as_str(),
AdminOperation(&TABLE_MAINTENANCE_QUARANTINE_HANDLER),
)?;
r.insert(
Method::GET,
format!("{prefix}/{{warehouse}}/namespaces/{{namespace}}/tables/{{table}}/catalog/export").as_str(),
@@ -5283,6 +5290,28 @@ impl Operation for HeartbeatTableMaintenanceJobHandler {
}
}
pub struct TableMaintenanceQuarantineHandler {}
#[async_trait::async_trait]
impl Operation for TableMaintenanceQuarantineHandler {
async fn call(&self, req: S3Request<Body>, params: Params<'_, '_>) -> S3Result<S3Response<(StatusCode, Body)>> {
let warehouse = warehouse_from_params(&params)?;
let namespace = namespace_from_params(&params)?;
let table = table_name_from_params(&params)?;
let job = job_id_from_params(&params)?;
let resource = TableCatalogResource::table(&warehouse, &namespace, &table);
authorize_table_catalog_resource_request(&req, &resource, AdminAction::RunTableMaintenanceAction).await?;
ensure_table_bucket_enabled(&warehouse).await?;
let request = read_json_body::<crate::table_catalog::TableMaintenanceQuarantineOperationRequest>(req.input).await?;
let store = table_catalog_store()?;
let response = store
.apply_table_maintenance_quarantine_operation(&warehouse, &namespace.public_name(), &table, &job, request)
.await
.map_err(catalog_store_error)?;
build_json_response(StatusCode::OK, &response)
}
}
pub struct ExportTableCatalogHandler {}
#[async_trait::async_trait]
@@ -5669,6 +5698,7 @@ mod tests {
("PutTableMaintenanceConfigHandler", "AdminAction::SetTableLifecycleAction"),
("GetTableMaintenanceJobHandler", "AdminAction::GetTableLifecycleAction"),
("GetTableMaintenanceSchedulerHandler", "AdminAction::GetTableLifecycleAction"),
("TableMaintenanceQuarantineHandler", "AdminAction::RunTableMaintenanceAction"),
("ExportTableCatalogHandler", "AdminAction::GetTableMetadataAction"),
("ImportTableCatalogHandler", "AdminAction::RegisterTableAction"),
("ExternalCatalogBridgeHandler", "AdminAction::GetTableMetadataAction"),
@@ -5723,6 +5753,7 @@ mod tests {
("PutTableMaintenanceConfigHandler", "AdminAction::SetTableLifecycleAction"),
("GetTableMaintenanceJobHandler", "AdminAction::GetTableLifecycleAction"),
("GetTableMaintenanceSchedulerHandler", "AdminAction::GetTableLifecycleAction"),
("TableMaintenanceQuarantineHandler", "AdminAction::RunTableMaintenanceAction"),
("ExportTableCatalogHandler", "AdminAction::GetTableMetadataAction"),
("ImportTableCatalogHandler", "AdminAction::RegisterTableAction"),
("ExternalCatalogBridgeHandler", "AdminAction::GetTableMetadataAction"),
@@ -5781,6 +5812,7 @@ mod tests {
"GetTableMaintenanceSchedulerHandler",
"RunTableMaintenanceWorkerHandler",
"HeartbeatTableMaintenanceJobHandler",
"TableMaintenanceQuarantineHandler",
"ExportTableCatalogHandler",
"ImportTableCatalogHandler",
"ExternalCatalogBridgeHandler",
@@ -5888,6 +5920,7 @@ mod tests {
let _: &PutTableMaintenanceConfigHandler = &PUT_TABLE_MAINTENANCE_CONFIG_HANDLER;
let _: &GetTableMaintenanceJobHandler = &GET_TABLE_MAINTENANCE_JOB_HANDLER;
let _: &GetTableMaintenanceSchedulerHandler = &GET_TABLE_MAINTENANCE_SCHEDULER_HANDLER;
let _: &TableMaintenanceQuarantineHandler = &TABLE_MAINTENANCE_QUARANTINE_HANDLER;
let _: &ExportTableCatalogHandler = &EXPORT_TABLE_CATALOG_HANDLER;
let _: &ImportTableCatalogHandler = &IMPORT_TABLE_CATALOG_HANDLER;
let _: &ExternalCatalogBridgeHandler = &EXTERNAL_CATALOG_BRIDGE_HANDLER;
@@ -5928,6 +5961,7 @@ mod tests {
assert_operation::<GetTableMaintenanceSchedulerHandler>();
assert_operation::<RunTableMaintenanceWorkerHandler>();
assert_operation::<HeartbeatTableMaintenanceJobHandler>();
assert_operation::<TableMaintenanceQuarantineHandler>();
assert_operation::<ExportTableCatalogHandler>();
assert_operation::<ImportTableCatalogHandler>();
assert_operation::<ExternalCatalogBridgeHandler>();
+23 -1
View File
@@ -847,6 +847,12 @@ pub const ADMIN_ROUTE_POLICY_SPECS: &[AdminRouteSpec] = &[
RUN_TABLE_MAINTENANCE,
RouteRiskLevel::High,
),
admin(
HttpMethod::Post,
"/iceberg/v1/{warehouse}/namespaces/{namespace}/tables/{table}/maintenance/jobs/{job}/quarantine",
RUN_TABLE_MAINTENANCE,
RouteRiskLevel::High,
),
admin(
HttpMethod::Get,
"/iceberg/v1/{warehouse}/namespaces/{namespace}/tables/{table}/catalog/export",
@@ -1100,6 +1106,12 @@ pub const ADMIN_ROUTE_POLICY_SPECS: &[AdminRouteSpec] = &[
RUN_TABLE_MAINTENANCE,
RouteRiskLevel::High,
),
admin(
HttpMethod::Post,
"/_iceberg/v1/{warehouse}/namespaces/{namespace}/tables/{table}/maintenance/jobs/{job}/quarantine",
RUN_TABLE_MAINTENANCE,
RouteRiskLevel::High,
),
admin(
HttpMethod::Get,
"/_iceberg/v1/{warehouse}/namespaces/{namespace}/tables/{table}/catalog/export",
@@ -1330,7 +1342,7 @@ mod tests {
let table_specs = ADMIN_ROUTE_POLICY_SPECS
.iter()
.filter(|spec| spec.path().starts_with("/iceberg/v1") || spec.path().starts_with("/_iceberg/v1"));
assert_eq!(table_specs.count(), 86);
assert_eq!(table_specs.count(), 88);
assert_action(HttpMethod::Put, "/iceberg/v1/buckets/{warehouse}", SET_TABLE_BUCKET);
assert_action(HttpMethod::Get, "/_iceberg/v1/buckets/{warehouse}", GET_TABLE_BUCKET);
assert_action(HttpMethod::Get, "/iceberg/v1/{warehouse}/namespaces", GET_TABLE_NAMESPACE);
@@ -1475,6 +1487,16 @@ mod tests {
"/_iceberg/v1/{warehouse}/namespaces/{namespace}/tables/{table}/maintenance/jobs/{job}/heartbeat",
RUN_TABLE_MAINTENANCE,
);
assert_action(
HttpMethod::Post,
"/iceberg/v1/{warehouse}/namespaces/{namespace}/tables/{table}/maintenance/jobs/{job}/quarantine",
RUN_TABLE_MAINTENANCE,
);
assert_action(
HttpMethod::Post,
"/_iceberg/v1/{warehouse}/namespaces/{namespace}/tables/{table}/maintenance/jobs/{job}/quarantine",
RUN_TABLE_MAINTENANCE,
);
assert_action(HttpMethod::Get, "/iceberg/v1/{warehouse}/catalog/migration", GET_TABLE_CATALOG);
assert_action(HttpMethod::Get, "/_iceberg/v1/{warehouse}/catalog/migration", GET_TABLE_CATALOG);
assert_action(
@@ -434,6 +434,11 @@ fn expected_admin_route_matrix() -> Vec<RouteMatrixEntry> {
"/{warehouse}/namespaces/{namespace}/tables/{table}/maintenance/jobs/{job}/heartbeat",
"/analytics/namespaces/sales/tables/orders/maintenance/jobs/job-1/heartbeat",
),
table_route_sample(
Method::POST,
"/{warehouse}/namespaces/{namespace}/tables/{table}/maintenance/jobs/{job}/quarantine",
"/analytics/namespaces/sales/tables/orders/maintenance/jobs/job-1/quarantine",
),
table_route_sample(
Method::GET,
"/{warehouse}/namespaces/{namespace}/tables/{table}/catalog/export",
@@ -613,6 +618,11 @@ fn expected_admin_route_matrix() -> Vec<RouteMatrixEntry> {
"/{warehouse}/namespaces/{namespace}/tables/{table}/maintenance/jobs/{job}/heartbeat",
"/analytics/namespaces/sales/tables/orders/maintenance/jobs/job-1/heartbeat",
),
compat_table_route_sample(
Method::POST,
"/{warehouse}/namespaces/{namespace}/tables/{table}/maintenance/jobs/{job}/quarantine",
"/analytics/namespaces/sales/tables/orders/maintenance/jobs/job-1/quarantine",
),
compat_table_route_sample(
Method::GET,
"/{warehouse}/namespaces/{namespace}/tables/{table}/catalog/export",
@@ -864,6 +874,11 @@ fn test_register_routes_cover_representative_admin_paths() {
Method::POST,
&table_catalog_path("/analytics/namespaces/sales/tables/orders/maintenance/jobs/job-1/heartbeat"),
);
assert_route(
&router,
Method::POST,
&table_catalog_path("/analytics/namespaces/sales/tables/orders/maintenance/jobs/job-1/quarantine"),
);
assert_route(
&router,
Method::GET,
@@ -1016,6 +1031,11 @@ fn test_register_routes_cover_representative_admin_paths() {
Method::POST,
&compat_table_catalog_path("/analytics/namespaces/sales/tables/orders/maintenance/jobs/job-1/heartbeat"),
);
assert_route(
&router,
Method::POST,
&compat_table_catalog_path("/analytics/namespaces/sales/tables/orders/maintenance/jobs/job-1/quarantine"),
);
assert_route(
&router,
Method::GET,
+374
View File
@@ -552,6 +552,30 @@ pub(crate) struct TableMaintenanceSchedulerQuarantineBoundary {
pub source_job_id: Option<String>,
}
#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)]
#[serde(rename_all = "SCREAMING_SNAKE_CASE")]
pub(crate) enum TableMaintenanceQuarantineAction {
Inspect,
Release,
Retry,
Abandon,
}
#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)]
#[serde(deny_unknown_fields)]
pub(crate) struct TableMaintenanceQuarantineOperationRequest {
pub action: TableMaintenanceQuarantineAction,
#[serde(default)]
pub reason: Option<String>,
}
#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)]
pub(crate) struct TableMaintenanceQuarantineOperationResult {
pub action: TableMaintenanceQuarantineAction,
pub report: TableMetadataMaintenanceReport,
pub scheduler: TableMaintenanceSchedulerReport,
}
#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)]
pub(crate) struct TableMaintenanceSchedulerJobSummary {
pub job_id: String,
@@ -3997,6 +4021,90 @@ where
})
}
pub(crate) async fn apply_table_maintenance_quarantine_operation(
&self,
table_bucket: &str,
namespace: &str,
table: &str,
job_id: &str,
request: TableMaintenanceQuarantineOperationRequest,
) -> TableCatalogStoreResult<TableMaintenanceQuarantineOperationResult> {
let action = request.action.clone();
let namespace = parse_namespace_for_store(namespace)?;
let table = parse_table_for_store(table)?;
let table_path = self.paths.table_entry_path(table_bucket, &namespace, &table);
let namespace_name = namespace.public_name();
let table_name = table.as_str().to_string();
let report = if matches!(action, TableMaintenanceQuarantineAction::Inspect) {
self.get_table_metadata_maintenance_report(table_bucket, &namespace_name, &table_name, job_id)
.await?
.ok_or_else(|| {
TableCatalogStoreError::NotFound(format!(
"maintenance job {}/{}/{}/{}",
table_bucket, namespace_name, table_name, job_id
))
})?
} else {
let _guard = self.backend.acquire_write_lock(self.catalog_bucket(), &table_path).await?;
let Some(mut report) = self
.get_table_metadata_maintenance_report(table_bucket, &namespace_name, &table_name, MAINTENANCE_JOB_ALIAS_CURRENT)
.await?
else {
return Err(TableCatalogStoreError::NotFound(format!(
"maintenance job {}/{}/{}/{}",
table_bucket, namespace_name, table_name, job_id
)));
};
if report.job.job_id != job_id {
return Err(TableCatalogStoreError::Conflict("maintenance job is not current".to_string()));
}
if !matches!(report.job.status, TableMetadataMaintenanceJobStatus::Failed) {
return Err(TableCatalogStoreError::Conflict(
"maintenance quarantine operation requires a failed job".to_string(),
));
}
if !report.job.quarantine_enabled || report.job.quarantined_object_count == 0 {
return Err(TableCatalogStoreError::Conflict(
"maintenance job has no active quarantine boundary".to_string(),
));
}
report.job.quarantined_object_count = 0;
match &action {
TableMaintenanceQuarantineAction::Inspect => unreachable!("inspect branch handled before mutation"),
TableMaintenanceQuarantineAction::Release => {
report.job.failure_reason =
Some(table_maintenance_quarantine_operator_reason("released", request.reason.as_deref()));
}
TableMaintenanceQuarantineAction::Retry => {
report.job.next_retry_after = None;
report.job.failure_reason = Some(table_maintenance_quarantine_operator_reason(
"released for retry",
request.reason.as_deref(),
));
}
TableMaintenanceQuarantineAction::Abandon => {
report.job.next_retry_after = None;
report.job.failure_reason =
Some(table_maintenance_quarantine_operator_reason("abandoned", request.reason.as_deref()));
}
}
refresh_table_maintenance_report_recommended_actions(&mut report);
self.put_table_metadata_maintenance_report(&report).await?;
report
};
let scheduler = self
.get_table_maintenance_scheduler_report(table_bucket, &namespace_name, &table_name)
.await?;
Ok(TableMaintenanceQuarantineOperationResult {
action,
report,
scheduler,
})
}
async fn list_table_metadata_maintenance_audit_reports(
&self,
table_bucket: &str,
@@ -6254,6 +6362,24 @@ where
}
}
pub(crate) async fn apply_table_maintenance_quarantine_operation(
&self,
table_bucket: &str,
namespace: &str,
table: &str,
job_id: &str,
request: TableMaintenanceQuarantineOperationRequest,
) -> TableCatalogStoreResult<TableMaintenanceQuarantineOperationResult> {
match self {
Self::ObjectBacked(store) => {
store
.apply_table_maintenance_quarantine_operation(table_bucket, namespace, table, job_id, request)
.await
}
Self::DurableStrong(_) => Err(Self::unsupported_for_durable_strong("table maintenance quarantine")),
}
}
pub(crate) async fn run_table_metadata_maintenance_worker_once(
&self,
table_bucket: &str,
@@ -8815,6 +8941,14 @@ fn parse_maintenance_timestamp(timestamp: &str) -> Option<OffsetDateTime> {
OffsetDateTime::parse(timestamp, &time::format_description::well_known::Rfc3339).ok()
}
fn table_maintenance_quarantine_operator_reason(action: &str, reason: Option<&str>) -> String {
let reason = reason.map(str::trim).filter(|reason| !reason.is_empty());
match reason {
Some(reason) => format!("maintenance quarantine {action} by operator: {reason}"),
None => format!("maintenance quarantine {action} by operator"),
}
}
fn table_maintenance_recommended_actions(job: &TableMetadataMaintenanceJob) -> Vec<TableMaintenanceRecommendedAction> {
let mut actions = Vec::new();
match job.status {
@@ -10842,6 +10976,58 @@ mod tests {
store.backend.reset_call_counts().await;
}
async fn seed_quarantined_table_maintenance(
store: &ObjectTableCatalogStore<TestCatalogObjectBackend>,
backend: &TestCatalogObjectBackend,
bucket: &str,
now: OffsetDateTime,
next_retry_after: Option<OffsetDateTime>,
) -> (Namespace, IdentifierSegment, TableMetadataMaintenanceReport) {
let namespace = Namespace::parse("sales").expect("namespace should parse");
let table = IdentifierSegment::parse("orders").expect("table should parse");
let current = default_table_metadata_file_path(&namespace, &table, "00002.metadata.json");
seed_table_for_metadata_maintenance(store, bucket, &namespace, &table, current.clone()).await;
backend
.seed_object(bucket, &current, br#"{"metadata-log":[]}"#.to_vec())
.await;
store
.put_table_maintenance_config(
bucket,
"sales",
"orders",
TableMaintenanceConfig {
version: TABLE_MAINTENANCE_CONFIG_VERSION,
background_enabled: true,
max_retry_attempts: 2,
retry_initial_backoff_seconds: 60,
retry_max_backoff_seconds: 300,
quarantine_enabled: true,
quarantine_retention_seconds: 86_400,
..Default::default()
},
)
.await
.expect("background maintenance config should persist");
let mut failed = store
.plan_table_metadata_maintenance(bucket, "sales", "orders", 0)
.await
.expect("maintenance report should be planned");
failed.job.status = TableMetadataMaintenanceJobStatus::Failed;
failed.job.failure_reason = Some("quarantine retained failed cleanup candidates".to_string());
failed.job.max_retry_attempts = 2;
failed.job.next_retry_after = next_retry_after.map(maintenance_timestamp);
failed.job.quarantine_enabled = true;
failed.job.quarantine_retention_seconds = 86_400;
failed.job.quarantined_object_count = 2;
failed.job.finished_at = Some(maintenance_timestamp(now - Duration::seconds(10)));
store
.put_table_metadata_maintenance_report(&failed)
.await
.expect("failed maintenance report should be seeded");
(namespace, table, failed)
}
#[tokio::test]
async fn object_table_catalog_store_writes_catalog_entries_to_internal_meta_bucket() {
let backend = TestCatalogObjectBackend::default();
@@ -12525,6 +12711,194 @@ mod tests {
);
}
#[tokio::test]
async fn maintenance_quarantine_retry_clears_boundary_and_unblocks_scheduler() {
let backend = TestCatalogObjectBackend::default();
let store = ObjectTableCatalogStore::new(backend.clone());
let bucket = "analytics";
let now = OffsetDateTime::now_utc();
let (_namespace, _table, failed) =
seed_quarantined_table_maintenance(&store, &backend, bucket, now, Some(now + Duration::seconds(300))).await;
let result = store
.apply_table_maintenance_quarantine_operation(
bucket,
"sales",
"orders",
&failed.job.job_id,
TableMaintenanceQuarantineOperationRequest {
action: TableMaintenanceQuarantineAction::Retry,
reason: Some("operator reviewed retained candidates".to_string()),
},
)
.await
.expect("quarantine retry should update the current maintenance job");
assert_eq!(result.action, TableMaintenanceQuarantineAction::Retry);
assert_eq!(result.report.job.job_id, failed.job.job_id);
assert_eq!(result.report.job.quarantined_object_count, 0);
assert!(result.report.job.next_retry_after.is_none());
assert!(
!result
.report
.job
.recommended_actions
.contains(&TableMaintenanceRecommendedAction::ReviewQuarantine)
);
assert_eq!(result.scheduler.status, TableMaintenanceSchedulerStatus::Ready);
assert!(!result.scheduler.quarantine.active);
let current_report = store
.get_table_metadata_maintenance_report(bucket, "sales", "orders", MAINTENANCE_JOB_ALIAS_CURRENT)
.await
.expect("current maintenance report should load")
.expect("current maintenance report should exist");
assert_eq!(current_report.job.job_id, failed.job.job_id);
assert_eq!(current_report.job.quarantined_object_count, 0);
}
#[tokio::test]
async fn maintenance_quarantine_inspect_reports_without_mutating_current_job() {
let backend = TestCatalogObjectBackend::default();
let store = ObjectTableCatalogStore::new(backend.clone());
let bucket = "analytics";
let now = OffsetDateTime::now_utc();
let (_namespace, _table, failed) =
seed_quarantined_table_maintenance(&store, &backend, bucket, now, Some(now + Duration::seconds(300))).await;
let result = store
.apply_table_maintenance_quarantine_operation(
bucket,
"sales",
"orders",
&failed.job.job_id,
TableMaintenanceQuarantineOperationRequest {
action: TableMaintenanceQuarantineAction::Inspect,
reason: Some("ignored for inspect".to_string()),
},
)
.await
.expect("quarantine inspect should load the maintenance job");
assert_eq!(result.action, TableMaintenanceQuarantineAction::Inspect);
assert_eq!(result.report.job.job_id, failed.job.job_id);
assert_eq!(result.report.job.quarantined_object_count, 2);
let expected_retry_after = maintenance_timestamp(now + Duration::seconds(300));
assert_eq!(result.report.job.next_retry_after.as_deref(), Some(expected_retry_after.as_str()));
assert_eq!(result.scheduler.status, TableMaintenanceSchedulerStatus::RetryDeferred);
let current_report = store
.get_table_metadata_maintenance_report(bucket, "sales", "orders", MAINTENANCE_JOB_ALIAS_CURRENT)
.await
.expect("current maintenance report should load")
.expect("current maintenance report should exist");
assert_eq!(current_report.job.quarantined_object_count, 2);
}
#[tokio::test]
async fn maintenance_quarantine_release_preserves_retry_deferral() {
let backend = TestCatalogObjectBackend::default();
let store = ObjectTableCatalogStore::new(backend.clone());
let bucket = "analytics";
let now = OffsetDateTime::now_utc();
let (_namespace, _table, failed) =
seed_quarantined_table_maintenance(&store, &backend, bucket, now, Some(now + Duration::seconds(300))).await;
let result = store
.apply_table_maintenance_quarantine_operation(
bucket,
"sales",
"orders",
&failed.job.job_id,
TableMaintenanceQuarantineOperationRequest {
action: TableMaintenanceQuarantineAction::Release,
reason: Some("objects retained for later retry".to_string()),
},
)
.await
.expect("quarantine release should update the current maintenance job");
assert_eq!(result.action, TableMaintenanceQuarantineAction::Release);
assert_eq!(result.report.job.quarantined_object_count, 0);
let expected_retry_after = maintenance_timestamp(now + Duration::seconds(300));
assert_eq!(result.report.job.next_retry_after.as_deref(), Some(expected_retry_after.as_str()));
assert_eq!(result.scheduler.status, TableMaintenanceSchedulerStatus::RetryDeferred);
assert!(result.report.job.failure_reason.as_deref().is_some_and(|reason| {
reason.contains("maintenance quarantine released by operator") && reason.contains("objects retained for later retry")
}));
}
#[tokio::test]
async fn maintenance_quarantine_abandon_clears_boundary_and_retry() {
let backend = TestCatalogObjectBackend::default();
let store = ObjectTableCatalogStore::new(backend.clone());
let bucket = "analytics";
let now = OffsetDateTime::now_utc();
let (_namespace, _table, failed) =
seed_quarantined_table_maintenance(&store, &backend, bucket, now, Some(now + Duration::seconds(300))).await;
let result = store
.apply_table_maintenance_quarantine_operation(
bucket,
"sales",
"orders",
&failed.job.job_id,
TableMaintenanceQuarantineOperationRequest {
action: TableMaintenanceQuarantineAction::Abandon,
reason: Some("operator accepted retained objects".to_string()),
},
)
.await
.expect("quarantine abandon should update the current maintenance job");
assert_eq!(result.action, TableMaintenanceQuarantineAction::Abandon);
assert_eq!(result.report.job.quarantined_object_count, 0);
assert!(result.report.job.next_retry_after.is_none());
assert_eq!(result.scheduler.status, TableMaintenanceSchedulerStatus::Ready);
assert!(result.report.job.failure_reason.as_deref().is_some_and(|reason| {
reason.contains("maintenance quarantine abandoned by operator")
&& reason.contains("operator accepted retained objects")
}));
}
#[tokio::test]
async fn maintenance_quarantine_rejects_mutating_non_current_job() {
let backend = TestCatalogObjectBackend::default();
let store = ObjectTableCatalogStore::new(backend.clone());
let bucket = "analytics";
let now = OffsetDateTime::now_utc();
let (_namespace, _table, old_failed) =
seed_quarantined_table_maintenance(&store, &backend, bucket, now, Some(now + Duration::seconds(300))).await;
let mut current_failed = store
.plan_table_metadata_maintenance(bucket, "sales", "orders", 0)
.await
.expect("maintenance report should be planned");
current_failed.job.status = TableMetadataMaintenanceJobStatus::Failed;
current_failed.job.quarantine_enabled = true;
current_failed.job.quarantine_retention_seconds = 86_400;
current_failed.job.quarantined_object_count = 1;
store
.put_table_metadata_maintenance_report(&current_failed)
.await
.expect("new current maintenance report should be seeded");
let error = store
.apply_table_maintenance_quarantine_operation(
bucket,
"sales",
"orders",
&old_failed.job.job_id,
TableMaintenanceQuarantineOperationRequest {
action: TableMaintenanceQuarantineAction::Release,
reason: None,
},
)
.await
.expect_err("mutating a non-current quarantine job should fail");
assert_eq!(error, TableCatalogStoreError::Conflict("maintenance job is not current".to_string()));
}
#[tokio::test]
async fn maintenance_worker_run_defers_until_retry_after() {
let backend = TestCatalogObjectBackend::default();
+3 -2
View File
@@ -232,7 +232,8 @@ The smoke test also probes catalog-backed advanced Iceberg surfaces:
drop routes with persisted view metadata and view-scoped authorization
- metadata maintenance supports safe dry-run planning, controlled worker
execution checks, and a scheduler status report with disabled, paused,
backpressure, retry, quarantine, and audit-timeline state
backpressure, retry, quarantine, and audit-timeline state; quarantined jobs
can be inspected, released, retried, or abandoned through an operator endpoint
- catalog diagnostics exposes the table recovery and consistency state used by
operators
- catalog export and diagnostics expose the current catalog backing manifest,
@@ -335,7 +336,7 @@ Unsupported behavior is documented instead of hidden behind internal errors. The
current unsupported inventory is:
- credential vending: automated after table bootstrap with exact-prefix validation and a data-plane scope probe; full no-long-term-data-credential bootstrap is not claimed
- background maintenance worker: controlled run-once, heartbeat, and scheduler status endpoints are registered; disabled/paused/backpressure/retry/quarantine/audit-timeline state is machine-readable; continuous in-process scheduling is not claimed
- background maintenance worker: controlled run-once, heartbeat, quarantine operation, and scheduler status endpoints are registered; disabled/paused/backpressure/retry/quarantine/audit-timeline state is machine-readable; continuous in-process scheduling is not claimed
- manifest/data reachability cleanup: metadata maintenance reads manifest-list and manifest Avro references, reports manifest/data/delete reachability, and deletes only unreferenced table objects that pass the safety window
- snapshot expiration dry-run planning and manual catalog commit: supported through metadata maintenance reports
- automatic maintenance scheduling: external scheduler hook supported through the worker run endpoint and scheduler status report; built-in periodic scheduling is not claimed
+10 -1
View File
@@ -183,7 +183,7 @@ UNSUPPORTED_INVENTORY: list[dict[str, str]] = [
"status": "controlled-run-once-supported",
"roadmap_area": "maintenance-worker",
"catalog_endpoint": "GET /v1/{prefix}/namespaces/{namespace}/tables/{table}/maintenance/scheduler",
"expected_behavior": "background-enabled maintenance can be driven by the worker run endpoint and inspected through the scheduler status endpoint with disabled/paused state, current-job backpressure, retry deferral, quarantine boundary, audit timeline, lease expiry recovery, and heartbeat updates; built-in periodic scheduling is not claimed",
"expected_behavior": "background-enabled maintenance can be driven by the worker run endpoint and inspected through the scheduler status endpoint with disabled/paused state, current-job backpressure, retry deferral, quarantine boundary, audit timeline, lease expiry recovery, heartbeat updates, and operator quarantine inspect/release/retry/abandon actions; built-in periodic scheduling is not claimed",
},
{
"capability": "manifest-data-reachability-cleanup",
@@ -1066,6 +1066,15 @@ def run_maintenance_probe(args: argparse.Namespace, deps: RuntimeDeps) -> None:
if not isinstance(job, dict) or not job.get("job-id"):
raise RuntimeError("maintenance metadata endpoint did not return a job id")
signed_rest_request(args, deps, "GET", table_endpoint_path(args, f"/maintenance/jobs/{job['job-id']}"))
quarantine = signed_rest_request(
args,
deps,
"POST",
table_endpoint_path(args, f"/maintenance/jobs/{job['job-id']}/quarantine"),
{"action": "INSPECT"},
)
if quarantine.get("action") != "INSPECT" or not isinstance(quarantine.get("report"), dict):
raise RuntimeError("maintenance quarantine endpoint did not return an inspection report")
scheduler = signed_rest_request(args, deps, "GET", scheduler_path)
expected_scheduler_statuses = {"READY", "DISABLED", "PAUSED", "BACKPRESSURED", "RETRY_DEFERRED", "QUARANTINED"}
if scheduler.get("status") not in expected_scheduler_statuses:
@@ -294,6 +294,7 @@ class PyIcebergSmokeConfigTest(unittest.TestCase):
config_path = pyiceberg_smoke.table_endpoint_path(args, "/maintenance/config")
maintenance_path = pyiceberg_smoke.table_endpoint_path(args, "/maintenance/metadata")
job_path = pyiceberg_smoke.table_endpoint_path(args, "/maintenance/jobs/job-1")
quarantine_path = pyiceberg_smoke.table_endpoint_path(args, "/maintenance/jobs/job-1/quarantine")
scheduler_path = pyiceberg_smoke.table_endpoint_path(args, "/maintenance/scheduler")
worker_path = pyiceberg_smoke.table_endpoint_path(args, "/maintenance/worker/run")
@@ -312,6 +313,8 @@ class PyIcebergSmokeConfigTest(unittest.TestCase):
return {"job": {"job-id": "job-1"}}
if (method, path) == ("GET", job_path):
return {"job": {"job-id": "job-1", "status": "SUCCESSFUL"}}
if (method, path) == ("POST", quarantine_path):
return {"action": "INSPECT", "report": {"job": {"job-id": "job-1"}}}
if (method, path) == ("GET", scheduler_path):
return {"status": "DISABLED", "audit_timeline": [{"job_id": "job-1"}]}
if (method, path) == ("POST", worker_path):
@@ -333,6 +336,7 @@ class PyIcebergSmokeConfigTest(unittest.TestCase):
view_path = pyiceberg_smoke.view_endpoint_path(args)
config_path = pyiceberg_smoke.table_endpoint_path(args, "/maintenance/config")
maintenance_path = pyiceberg_smoke.table_endpoint_path(args, "/maintenance/metadata")
quarantine_path = pyiceberg_smoke.table_endpoint_path(args, "/maintenance/jobs/job-1/quarantine")
scheduler_path = pyiceberg_smoke.table_endpoint_path(args, "/maintenance/scheduler")
worker_path = pyiceberg_smoke.table_endpoint_path(args, "/maintenance/worker/run")
diagnostics_path = pyiceberg_smoke.table_endpoint_path(args, "/catalog/diagnostics")
@@ -373,6 +377,8 @@ class PyIcebergSmokeConfigTest(unittest.TestCase):
return {"job": {"job-id": "job-1"}}
if (method, path) == ("GET", pyiceberg_smoke.table_endpoint_path(args, "/maintenance/jobs/job-1")):
return {"job": {"job-id": "job-1", "status": "SUCCESSFUL"}}
if (method, path) == ("POST", quarantine_path):
return {"action": "INSPECT", "report": {"job": {"job-id": "job-1"}}}
if (method, path) == ("GET", scheduler_path):
return {"status": "DISABLED", "audit_timeline": [{"job_id": "job-1"}]}
if (method, path) == ("POST", worker_path):
@@ -405,6 +411,7 @@ class PyIcebergSmokeConfigTest(unittest.TestCase):
self.assertIn(("GET", metadata_location_path, None), calls)
self.assertIn(("GET", diagnostics_path, None), calls)
self.assertIn(("GET", scheduler_path, None), calls)
self.assertIn(("POST", quarantine_path, {"action": "INSPECT"}), calls)
self.assertIn(("POST", worker_path, {}), calls)
def test_table_ref_probe_force_deletes_smoke_ref_after_validation_failure(self) -> None: