From d4c28ee3178a25941f363a34e45a08cdae28fbd3 Mon Sep 17 00:00:00 2001
From: Avi Kivity <avi@scylladb.com>
Date: Mon, 6 Apr 2026 14:07:48 +0300
Subject: [PATCH] Merge 'service_levels: mark v2 migration complete on empty
 legacy table' from Alex Dathskovsky

During raft-topology upgrade in 2026.1, service_level_controller::migrate_to_v2() returns early when system_distributed.service_levels is empty. This skips the service_level_version = 2 write, so the cluster is never marked as upgraded to service levels v2 even though there is no data to migrate. Subsequent upgrades may then fail the startup check which requires service_level_version == 2.
Remove the early return and let the migration commit the version marker even when there are no legacy service levels rows to copy.

Fixes: https://scylladb.atlassian.net/browse/SCYLLADB-1198

backport: should be backported to all versions that can be upgraded to 2026.2

Closes scylladb/scylladb#29333

* github.com:scylladb/scylladb:
  test/auth_cluster: cover empty legacy table in service level upgrade
  service_levels: mark v2 migration complete on empty legacy table

(cherry picked from commit 95e422db482b3441f8bdba6563a58d5463fa5060)

Closes scylladb/scylladb#29352
---
 service/qos/service_level_controller.cc       |  4 ----
 .../auth_cluster/test_raft_service_levels.py  | 21 +++++++++++++++++++
 2 files changed, 21 insertions(+), 4 deletions(-)

diff --git a/service/qos/service_level_controller.cc b/service/qos/service_level_controller.cc
index dc874b0142..33d1683bfd 100644
--- a/service/qos/service_level_controller.cc
+++ b/service/qos/service_level_controller.cc
@@ -880,10 +880,6 @@ future<> service_level_controller::migrate_to_v2(size_t nodes_count, db::system_
         qs,
         {},
         cql3::query_processor::cache_internal::no);
-    if (rows->empty()) {
-        co_return;
-    }
-    
 
     auto col_names = schema->all_columns() | std::views::transform([] (const auto& col) {return col.name_as_cql_string(); }) | std::ranges::to<std::vector<sstring>>();
     auto col_names_str = fmt::to_string(fmt::join(col_names, ", "));
diff --git a/test/cluster/auth_cluster/test_raft_service_levels.py b/test/cluster/auth_cluster/test_raft_service_levels.py
index 2e6efd8916..cfa55e451f 100644
--- a/test/cluster/auth_cluster/test_raft_service_levels.py
+++ b/test/cluster/auth_cluster/test_raft_service_levels.py
@@ -109,6 +109,27 @@ async def test_service_levels_upgrade(request, manager: ManagerClient, build_mod
     result_with_sl_v2 = await cql.run_async(f"SELECT service_level FROM system.service_levels_v2")
     assert set([sl.service_level for sl in result_with_sl_v2]) == set(sls + [sl_v2])
 
+@pytest.mark.asyncio
+async def test_service_levels_upgrade_with_empty_legacy_table(manager: ManagerClient):
+    cfg = {**auth_config, "force_gossip_topology_changes": True, "tablets_mode_for_new_keyspaces": "disabled"}
+
+    servers = [await manager.server_add(config=cfg)]
+    cfg.pop("force_gossip_topology_changes")
+    servers += [await manager.server_add(config=cfg) for _ in range(2)]
+
+    cql = manager.get_cql()
+    assert cql
+    hosts = await wait_for_cql_and_get_hosts(cql, servers, time.time() + 60)
+
+    rows = await cql.run_async("SELECT service_level FROM system_distributed.service_levels")
+    assert list(rows) == []
+
+    await manager.api.upgrade_to_raft_topology(hosts[0].address)
+    await asyncio.gather(*(wait_until_topology_upgrade_finishes(manager, h.address, time.time() + 60) for h in hosts))
+
+    sl_version = await cql.run_async("SELECT value FROM system.scylla_local WHERE key = 'service_level_version'")
+    assert sl_version[0].value == "2"
+
 @pytest.mark.asyncio
 async def test_service_levels_work_during_recovery(manager: ManagerClient):
     # FIXME: move this test to the Raft-based recovery procedure or remove it if unneeded.