diff --git a/service/strong_consistency/groups_manager.cc b/service/strong_consistency/groups_manager.cc
index 45501b7ec4..701e40a76a 100644
--- a/service/strong_consistency/groups_manager.cc
+++ b/service/strong_consistency/groups_manager.cc
@@ -137,6 +137,12 @@ future<> groups_manager::start_raft_group(global_tablet_id tablet,
 
     auto& persistence_ref = *storage;
     auto config = raft::server::configuration {
+        // Snapshotting is not implemented yet for strong consistency,
+        // so effectively disable periodic snapshotting.
+        // TODO: Revert after snapshots are implemented
+        .snapshot_threshold = std::numeric_limits<size_t>::max(),
+        .snapshot_threshold_log_size = 10 * 1024 * 1024, // 10MB
+        .max_log_size = 20 * 1024 * 1024, // 20MB
         .enable_forwarding = false,
         .on_background_error = [tablet, group_id](std::exception_ptr e) {
             on_internal_error(logger, 
diff --git a/service/strong_consistency/state_machine.cc b/service/strong_consistency/state_machine.cc
index e8e80d0214..0154df33b7 100644
--- a/service/strong_consistency/state_machine.cc
+++ b/service/strong_consistency/state_machine.cc
@@ -79,11 +79,16 @@ public:
     }
 
     future<raft::snapshot_id> take_snapshot() override {
-        throw std::runtime_error("take_snapshot() not implemented");
+        // Until snapshot transfer is fully implemented, return a fake ID
+        // and don't actually do anything. As long as we don't do snapshot
+        // transfers (attempting to do that throws an exception), we should
+        // be safe.
+        return make_ready_future<raft::snapshot_id>(raft::snapshot_id(utils::make_random_uuid()));
     }
 
     void drop_snapshot(raft::snapshot_id id) override {
-        throw std::runtime_error("drop_snapshot() not implemented");
+        // Taking a snapshot is a no-op, so dropping a snapshot is also a no-op.
+        (void) id;
     }
 
     future<> load_snapshot(raft::snapshot_id id) override {