Add timeline_import http endpoint (#11484)

The added `timleine_import` endpoint allows us to migrate safekeeper
timelines from control plane managed to storcon managed.
 
Part of #9011
This commit is contained in:
Arpad Müller
2025-04-11 16:10:27 +02:00
committed by GitHub
parent 88f01c1ca1
commit c66444ea15
3 changed files with 82 additions and 2 deletions

View File

@@ -7,7 +7,8 @@ use std::time::{Duration, Instant};
/// API (`/control/v1` prefix). Implemented by the server
/// in [`storage_controller::http`]
use serde::{Deserialize, Serialize};
use utils::id::{NodeId, TenantId};
use utils::id::{NodeId, TenantId, TimelineId};
use utils::lsn::Lsn;
use crate::models::{PageserverUtilization, ShardParameters, TenantConfig};
use crate::shard::{ShardStripeSize, TenantShardId};
@@ -499,6 +500,15 @@ pub struct SafekeeperSchedulingPolicyRequest {
pub scheduling_policy: SkSchedulingPolicy,
}
/// Import request for safekeeper timelines.
#[derive(Serialize, Deserialize, Clone)]
pub struct TimelineImportRequest {
pub tenant_id: TenantId,
pub timeline_id: TimelineId,
pub start_lsn: Lsn,
pub sk_set: Vec<NodeId>,
}
#[cfg(test)]
mod test {
use serde_json;

View File

@@ -22,6 +22,7 @@ use pageserver_api::controller_api::{
MetadataHealthListUnhealthyResponse, MetadataHealthUpdateRequest, MetadataHealthUpdateResponse,
NodeAvailability, NodeConfigureRequest, NodeRegisterRequest, SafekeeperSchedulingPolicyRequest,
ShardsPreferredAzsRequest, TenantCreateRequest, TenantPolicyRequest, TenantShardMigrateRequest,
TimelineImportRequest,
};
use pageserver_api::models::{
DetachBehavior, LsnLeaseRequest, TenantConfigPatchRequest, TenantConfigRequest,
@@ -1286,6 +1287,37 @@ async fn handle_tenant_import(req: Request<Body>) -> Result<Response<Body>, ApiE
)
}
async fn handle_timeline_import(req: Request<Body>) -> Result<Response<Body>, ApiError> {
let tenant_id: TenantId = parse_request_param(&req, "tenant_id")?;
let timeline_id: TimelineId = parse_request_param(&req, "timeline_id")?;
check_permissions(&req, Scope::PageServerApi)?;
maybe_rate_limit(&req, tenant_id).await;
let mut req = match maybe_forward(req).await {
ForwardOutcome::Forwarded(res) => {
return res;
}
ForwardOutcome::NotForwarded(req) => req,
};
let import_req = json_request::<TimelineImportRequest>(&mut req).await?;
let state = get_state(&req);
if import_req.tenant_id != tenant_id || import_req.timeline_id != timeline_id {
return Err(ApiError::BadRequest(anyhow::anyhow!(
"tenant id or timeline id mismatch: url={tenant_id}/{timeline_id}, body={}/{}",
import_req.tenant_id,
import_req.timeline_id
)));
}
json_response(
StatusCode::OK,
state.service.timeline_import(import_req).await?,
)
}
async fn handle_tenants_dump(req: Request<Body>) -> Result<Response<Body>, ApiError> {
check_permissions(&req, Scope::Admin)?;
@@ -1959,6 +1991,16 @@ pub fn make_router(
RequestName("debug_v1_tenant_locate"),
)
})
.post(
"/debug/v1/tenant/:tenant_id/timeline/:timeline_id/import",
|r| {
named_request_span(
r,
handle_timeline_import,
RequestName("debug_v1_timeline_import"),
)
},
)
.get("/debug/v1/scheduler", |r| {
named_request_span(r, handle_scheduler_dump, RequestName("debug_v1_scheduler"))
})

View File

@@ -12,13 +12,16 @@ use crate::persistence::{
use crate::safekeeper::Safekeeper;
use anyhow::Context;
use http_utils::error::ApiError;
use pageserver_api::controller_api::{SafekeeperDescribeResponse, SkSchedulingPolicy};
use pageserver_api::controller_api::{
SafekeeperDescribeResponse, SkSchedulingPolicy, TimelineImportRequest,
};
use pageserver_api::models::{self, SafekeeperInfo, SafekeepersInfo, TimelineInfo};
use safekeeper_api::membership::{MemberSet, SafekeeperId};
use tokio::task::JoinSet;
use tokio_util::sync::CancellationToken;
use utils::id::{NodeId, TenantId, TimelineId};
use utils::logging::SecretString;
use utils::lsn::Lsn;
use super::Service;
@@ -298,6 +301,31 @@ impl Service {
timeline_id,
})
}
/// Directly insert the timeline into the database without reconciling it with safekeepers.
///
/// Useful if the timeline already exists on the specified safekeepers,
/// but we want to make it storage controller managed.
pub(crate) async fn timeline_import(&self, req: TimelineImportRequest) -> Result<(), ApiError> {
let persistence = TimelinePersistence {
tenant_id: req.tenant_id.to_string(),
timeline_id: req.timeline_id.to_string(),
start_lsn: Lsn::INVALID.into(),
generation: 1,
sk_set: req.sk_set.iter().map(|sk_id| sk_id.0 as i64).collect(),
new_sk_set: None,
cplane_notified_generation: 1,
deleted_at: None,
};
let inserted = self.persistence.insert_timeline(persistence).await?;
if inserted {
tracing::info!("imported timeline into db");
} else {
tracing::info!("didn't import timeline into db, as it is already present in db");
}
Ok(())
}
/// Perform timeline deletion on safekeepers. Will return success: we persist the deletion into the reconciler.
pub(super) async fn tenant_timeline_delete_safekeepers(
self: &Arc<Self>,