rust/neon
1
0
Fork 0
mirror of https://github.com/neondatabase/neon.git synced 2026-05-24 00:20:37 +00:00
Code Issues Packages Projects Releases Wiki Activity

storcon: implement safekeeper_migrate_abort handler (#12705)

Browse Source 
## Problem
Right now if we commit a joint configuration to DB, there is no way
back. The only way to get the clean mconf is to continue the migration.
The RFC also described an abort mechanism, which allows to abort current
migration and revert mconf change. It might be needed if the migration
is stuck and cannot have any progress, e.g. if the sk we are migrating
to went down during the migration. This PR implements this abort
algorithm.

- Closes: https://databricks.atlassian.net/browse/LKB-899
- Closes: https://github.com/neondatabase/neon/issues/12549

## Summary of changes
- Implement `safekeeper_migrate_abort` handler with the algorithm
described in RFC
- Add `timeline-safekeeper-migrate-abort` subcommand to `storcon_cli`
- Add test for the migration abort algorithm.
This commit is contained in:
Dmitrii Kovalkov
2025-07-31 16:40:32 +04:00
committed by GitHub
parent df4e37b7cc
commit 312a74f11f
5 changed files with 266 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

28
storage_controller/src/http.rs
View File

@@ -644,6 +644,7 @@ async fn handle_tenant_timeline_safekeeper_migrate(
req: Request<Body>,
) -> Result<Response<Body>, ApiError> {
let tenant_id: TenantId = parse_request_param(&req, "tenant_id")?;
// TODO(diko): it's not PS operation, there should be a different permission scope.
check_permissions(&req, Scope::PageServerApi)?;
maybe_rate_limit(&req, tenant_id).await;
@@ -665,6 +666,23 @@ async fn handle_tenant_timeline_safekeeper_migrate(
json_response(StatusCode::OK, ())
}
async fn handle_tenant_timeline_safekeeper_migrate_abort(
service: Arc<Service>,
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")?;
// TODO(diko): it's not PS operation, there should be a different permission scope.
check_permissions(&req, Scope::PageServerApi)?;
maybe_rate_limit(&req, tenant_id).await;
service
.tenant_timeline_safekeeper_migrate_abort(tenant_id, timeline_id)
.await?;
json_response(StatusCode::OK, ())
}
async fn handle_tenant_timeline_lsn_lease(
service: Arc<Service>,
req: Request<Body>,
@@ -2611,6 +2629,16 @@ pub fn make_router(
)
},
)
.post(
"/v1/tenant/:tenant_id/timeline/:timeline_id/safekeeper_migrate_abort",
|r| {
tenant_service_handler(
r,
handle_tenant_timeline_safekeeper_migrate_abort,
RequestName("v1_tenant_timeline_safekeeper_migrate_abort"),
)
},
)
// LSN lease passthrough to all shards
.post(
"/v1/tenant/:tenant_id/timeline/:timeline_id/lsn_lease",