migration_manager: take group0 lock during raft snapshot taking

Group0 state machine access atomicity is guaranteed by a mutex in group0
client. A code that reads or writes the state needs to hold the log. To
transfer schema part of the snapshot we used existing "migration
request" verb which did not follow the rule. Fix the code to take group0
lock before accessing schema in case the verb is called as part of
group0 snapshot transfer.

Fixes #16821

service/migration_manager.cc

@@ -156,6 +156,10 @@ void migration_manager::init_messaging_service()
         auto features = self._feat.cluster_schema_features();
         auto& proxy = self._storage_proxy.container();
         auto& db = proxy.local().get_db();
+        semaphore_units<> guard;
+        if (options->group0_snapshot_transfer) {
+            guard = co_await self._group0_client.hold_read_apply_mutex(self._as);
+        }
         auto cm = co_await db::schema_tables::convert_schema_to_mutations(proxy, features);
         if (options->group0_snapshot_transfer) {
             cm.emplace_back(co_await db::system_keyspace::get_group0_history(db));