storage_service: register schema version observer before joining grou…

…p0 and starting gossiper

The schema version is updated by group0, so if group0 starts before
schema version observer is registered some updates may be missed. Since
the observer is used to update node's gossiper state the gossiper may
contain wrong schema version.

Fix by registering the observer before starting group0 and even before
starting gossiper to avoid a theoretical case that something may pull
schema after start of gossiping and before the observer is registered.

Fixes: scylladb#15078

Message-Id: <ZOYZWhEh6Zyb+FaN@scylladb.com>

service/storage_service.cc

@@ -2485,6 +2485,12 @@ future<> storage_service::join_token_ring(cdc::generation_service& cdc_gen_servi
         app_states.emplace(s.first, std::move(s.second));
     }
 
+    auto schema_change_announce = _db.local().observable_schema_version().observe([this] (table_schema_version schema_version) mutable {
+        _migration_manager.local().passive_announce(std::move(schema_version));
+    });
+
+    _listeners.emplace_back(make_lw_shared(std::move(schema_change_announce)));
+
     slogger.info("Starting up server gossip");
 
     auto generation_number = gms::generation_type(co_await _sys_ks.local().increment_and_get_generation());
@@ -2557,10 +2563,6 @@ future<> storage_service::join_token_ring(cdc::generation_service& cdc_gen_servi
         }
     } ();
 
-    auto schema_change_announce = _db.local().observable_schema_version().observe([this] (table_schema_version schema_version) mutable {
-        _migration_manager.local().passive_announce(std::move(schema_version));
-    });
-    _listeners.emplace_back(make_lw_shared(std::move(schema_change_announce)));
     co_await _gossiper.wait_for_gossip_to_settle();
     // TODO: Look at the group 0 upgrade state and use it to decide whether to attach or not
     if (!_raft_topology_change_enabled) {