rpc: add new join handshake verbs

The `join_node_request` and `join_node_response` RPCs are added:

- `join_node_request` is sent from the joining node to any node in the
  cluster. It contains some initial parameters that will be verified by
  the receiving node, or the topology coordinator - notably, it contains
  a list of cluster features supported by the joining node.
- `join_node_response` is sent from the topology coordinator to the
  joining node to tell it about the the outcome of the verification.

configure.py

@@ -1318,6 +1318,7 @@ idls = ['idl/gossip_digest.idl.hh',
         'idl/position_in_partition.idl.hh',
         'idl/experimental/broadcast_tables_lang.idl.hh',
         'idl/storage_service.idl.hh',
+        'idl/join_node.idl.hh',
         'idl/utils.idl.hh',
         ]
 

idl/join_node.idl.hh

@@ -0,0 +1,56 @@
+/*
+ * Copyright 2023-present ScyllaDB
+ */
+
+/*
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+
+namespace service {
+
+struct join_node_request_params {
+    raft::server_id host_id;
+    std::optional<raft::server_id> replaced_id;
+    std::vector<sstring> ignore_nodes;
+    sstring cluster_name;
+    sstring snitch_name;
+    sstring datacenter;
+    sstring rack;
+    sstring release_version;
+    uint32_t num_tokens;
+    uint32_t shard_count;
+    uint32_t ignore_msb;
+    std::vector<sstring> supported_features;
+};
+
+struct join_node_request_result {
+    struct ok {};
+    struct rejected {
+        sstring reason;
+    };
+
+    std::variant<
+        service::join_node_request_result::ok,
+        service::join_node_request_result::rejected
+    > result;
+};
+
+struct join_node_response_params {
+    struct accepted {};
+
+    struct rejected {
+        sstring reason;
+    };
+
+    std::variant<
+        service::join_node_response_params::accepted,
+        service::join_node_response_params::rejected
+    > response;
+};
+
+struct join_node_response_result {};
+
+verb join_node_request (raft::server_id dst_id, service::join_node_request_params) -> service::join_node_request_result;
+verb join_node_response (raft::server_id dst_id, service::join_node_response_params) -> service::join_node_response_result;
+
+}

message/messaging_service.cc

@@ -48,6 +48,7 @@
 #include "full_position.hh"
 #include "db/per_partition_rate_limit_info.hh"
 #include "service/topology_state_machine.hh"
+#include "service/raft/join_node.hh"
 #include "idl/consistency_level.dist.hh"
 #include "idl/tracing.dist.hh"
 #include "idl/result.dist.hh"
@@ -77,6 +78,7 @@
 #include "idl/per_partition_rate_limit_info.dist.hh"
 #include "idl/storage_proxy.dist.hh"
 #include "idl/storage_service.dist.hh"
+#include "idl/join_node.dist.hh"
 #include "message/rpc_protocol_impl.hh"
 #include "idl/consistency_level.dist.impl.hh"
 #include "idl/tracing.dist.impl.hh"
@@ -118,6 +120,7 @@
 #include "idl/forward_request.dist.hh"
 #include "idl/forward_request.dist.impl.hh"
 #include "idl/storage_service.dist.impl.hh"
+#include "idl/join_node.dist.impl.hh"
 
 namespace netw {
 
@@ -566,6 +569,8 @@ static constexpr unsigned do_get_rpc_client_idx(messaging_verb verb) {
     case messaging_verb::GROUP0_MODIFY_CONFIG:
     case messaging_verb::GET_GROUP0_UPGRADE_STATE:
     case messaging_verb::RAFT_TOPOLOGY_CMD:
+    case messaging_verb::JOIN_NODE_REQUEST:
+    case messaging_verb::JOIN_NODE_RESPONSE:
         // See comment above `TOPOLOGY_INDEPENDENT_IDX`.
         // DO NOT put any 'hot' (e.g. data path) verbs in this group,
         // only verbs which are 'rare' and 'cheap'.

message/messaging_service.hh

@@ -184,7 +184,9 @@ enum class messaging_verb : int32_t {
     RAFT_PULL_TOPOLOGY_SNAPSHOT = 65,
     TABLET_STREAM_DATA = 66,
     TABLET_CLEANUP = 67,
-    LAST = 68,
+    JOIN_NODE_REQUEST = 68,
+    JOIN_NODE_RESPONSE = 69,
+    LAST = 70,
 };
 
 } // namespace netw

service/raft/join_node.hh

@@ -0,0 +1,62 @@
+/*
+ * Copyright 2023-present ScyllaDB
+ */
+
+/*
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+
+#pragma once
+
+#include <set>
+#include <seastar/core/sstring.hh>
+#include "raft/raft.hh"
+
+namespace service {
+
+struct join_node_request_params {
+    raft::server_id host_id;
+    std::optional<raft::server_id> replaced_id;
+    std::vector<sstring> ignore_nodes;
+    sstring cluster_name;
+    sstring snitch_name;
+    sstring datacenter;
+    sstring rack;
+    sstring release_version;
+    uint32_t num_tokens;
+    uint32_t shard_count;
+    uint32_t ignore_msb;
+    std::vector<sstring> supported_features;
+};
+
+struct join_node_request_result {
+    // Request was successfully placed and will be processed
+    // by the topology coordinator.
+    struct ok {};
+
+    // The request was immediately rejected, most likely due to some
+    // parameters being incorrect or incompatible with the cluster.
+    struct rejected {
+        sstring reason;
+    };
+
+    std::variant<ok, rejected> result;
+};
+
+struct join_node_response_params {
+    // The topology coordinator accepts and wants to add the joining node
+    // to group 0 and to the cluster in general.
+    struct accepted {};
+
+    // The topology coordinator rejects the node, most likely due to some
+    // parameters being incorrect or incompatible with the cluster.
+    struct rejected {
+        sstring reason;
+    };
+
+    std::variant<accepted, rejected> response;
+};
+
+struct join_node_response_result {};
+
+}