/** * Copyright (c) 2013, Willem-Hendrik Thiart * Use of this source code is governed by a BSD-style license that can be * found in the LICENSE file. * * @file * @brief Implementation of a Raft server * @author Willem Thiart himself@willemthiart.com * @version 0.1 */ #include #include #include #include /* for varags */ #include #include "raft.h" #include "raft_log.h" #include "raft_private.h" static void __log(raft_server_t *me_, const char *fmt, ...) { char buf[1024]; va_list args; va_start(args, fmt); vsprintf(buf, fmt, args); #if 0 /* debugging */ raft_server_private_t* me = (void*)me_; printf("%d: %s\n", me->nodeid, buf); __FUNC_log(bto, src, buf); #endif } raft_server_t* raft_new() { raft_server_private_t* me; if (!(me = calloc(1, sizeof(raft_server_private_t)))) return NULL; me->current_term = 0; me->voted_for = -1; me->current_idx = 1; me->timeout_elapsed = 0; me->request_timeout = 200; me->election_timeout = 1000; me->log = log_new(); raft_set_state((void*)me, RAFT_STATE_FOLLOWER); return (void*)me; } void raft_set_callbacks(raft_server_t* me_, raft_cbs_t* funcs, void* udata) { raft_server_private_t* me = (void*)me_; memcpy(&me->cb, funcs, sizeof(raft_cbs_t)); me->udata = udata; } void raft_free(raft_server_t* me_) { raft_server_private_t* me = (void*)me_; log_free(me->log); free(me_); } void raft_election_start(raft_server_t* me_) { raft_server_private_t* me = (void*)me_; __log(me_, "election starting: %d %d, term: %d", me->election_timeout, me->timeout_elapsed, me->current_term); raft_become_candidate(me_); } void raft_become_leader(raft_server_t* me_) { raft_server_private_t* me = (void*)me_; int i; __log(me_, "becoming leader"); raft_set_state(me_, RAFT_STATE_LEADER); me->voted_for = -1; for (i = 0; i < me->num_nodes; i++) { if (me->nodeid == i) continue; raft_node_t* p = raft_get_node(me_, i); raft_node_set_next_idx(p, raft_get_current_idx(me_) + 1); raft_send_appendentries(me_, i); } } void raft_become_candidate(raft_server_t* me_) { raft_server_private_t* me = (void*)me_; int i; __log(me_, "becoming candidate"); memset(me->votes_for_me, 0, sizeof(int) * me->num_nodes); me->current_term += 1; raft_vote(me_, me->nodeid); raft_set_state(me_, RAFT_STATE_CANDIDATE); /* we need a random factor here to prevent simultaneous candidates */ me->timeout_elapsed = rand() % 500; for (i = 0; i < me->num_nodes; i++) { if (me->nodeid == i) continue; raft_send_requestvote(me_, i); } } void raft_become_follower(raft_server_t* me_) { raft_server_private_t* me = (void*)me_; __log(me_, "becoming follower"); raft_set_state(me_, RAFT_STATE_FOLLOWER); me->voted_for = -1; } int raft_periodic(raft_server_t* me_, int msec_since_last_period) { raft_server_private_t* me = (void*)me_; __log(me_, "periodic elapsed time: %d", me->timeout_elapsed); switch (me->state) { case RAFT_STATE_FOLLOWER: if (me->last_applied_idx < me->commit_idx) if (0 == raft_apply_entry(me_)) return 0; break; } me->timeout_elapsed += msec_since_last_period; if (me->state == RAFT_STATE_LEADER) { if (me->request_timeout <= me->timeout_elapsed) { raft_send_appendentries_all(me_); me->timeout_elapsed = 0; } } else if (me->election_timeout <= me->timeout_elapsed) raft_election_start(me_); return 1; } raft_entry_t* raft_get_entry_from_idx(raft_server_t* me_, int etyidx) { raft_server_private_t* me = (void*)me_; return log_get_from_idx(me->log, etyidx); } int raft_recv_appendentries_response(raft_server_t* me_, int node, msg_appendentries_response_t* r) { raft_server_private_t* me = (void*)me_; raft_node_t* p; __log(me_, "received appendentries response from: %d", node); p = raft_get_node(me_, node); if (1 == r->success) { int i; for (i = r->first_idx; i <= r->current_idx; i++) log_mark_node_has_committed(me->log, i); while (1) { raft_entry_t* e; e = log_get_from_idx(me->log, me->last_applied_idx + 1); /* majority has this */ if (e && me->num_nodes / 2 <= e->num_nodes) { if (0 == raft_apply_entry(me_)) break; } else break; } } else { /* If AppendEntries fails because of log inconsistency: decrement nextIndex and retry (§5.3) */ assert(0 <= raft_node_get_next_idx(p)); // TODO does this have test coverage? // TODO can jump back to where node is different instead of iterating raft_node_set_next_idx(p, raft_node_get_next_idx(p) - 1); raft_send_appendentries(me_, node); } return 1; } int raft_recv_appendentries( raft_server_t* me_, const int node, msg_appendentries_t* ae) { raft_server_private_t* me = (void*)me_; msg_appendentries_response_t r; me->timeout_elapsed = 0; __log(me_, "received appendentries from: %d", node); r.term = me->current_term; /* we've found a leader who is legitimate */ if (raft_is_leader(me_) && me->current_term <= ae->term) raft_become_follower(me_); /* 1. Reply false if term < currentTerm (§5.1) */ if (ae->term < me->current_term) { __log(me_, "AE term is less than current term"); r.success = 0; goto done; } #if 0 if (-1 != ae->prev_log_idx && ae->prev_log_idx < raft_get_current_idx(me_)) { __log(me_, "AE prev_idx is less than current idx"); r.success = 0; goto done; } #endif /* not the first appendentries we've received */ if (0 != ae->prev_log_idx) { raft_entry_t* e; if ((e = raft_get_entry_from_idx(me_, ae->prev_log_idx))) { /* 2. Reply false if log doesn’t contain an entry at prevLogIndex whose term matches prevLogTerm (§5.3) */ if (e->term != ae->prev_log_term) { __log(me_, "AE term doesn't match prev_idx"); r.success = 0; goto done; } /* 3. If an existing entry conflicts with a new one (same index but different terms), delete the existing entry and all that follow it (§5.3) */ raft_entry_t* e2; if ((e2 = raft_get_entry_from_idx(me_, ae->prev_log_idx + 1))) log_delete(me->log, ae->prev_log_idx + 1); } else { __log(me_, "AE no log at prev_idx"); r.success = 0; goto done; //assert(0); } } /* 5. If leaderCommit > commitIndex, set commitIndex = min(leaderCommit, last log index) */ if (raft_get_commit_idx(me_) < ae->leader_commit) { raft_entry_t* e; if ((e = log_peektail(me->log))) { raft_set_commit_idx(me_, e->id < ae->leader_commit ? e->id : ae->leader_commit); while (1 == raft_apply_entry(me_)) ; } } if (raft_is_candidate(me_)) raft_become_follower(me_); raft_set_current_term(me_, ae->term); int i; /* append all entries to log */ for (i = 0; i < ae->n_entries; i++) { msg_entry_t* cmd; raft_entry_t* c; cmd = &ae->entries[i]; /* TODO: replace malloc with mempoll/arena */ c = malloc(sizeof(raft_entry_t)); c->term = me->current_term; c->len = cmd->len; c->id = cmd->id; c->data = malloc(cmd->len); memcpy(c->data, cmd->data, cmd->len); if (0 == raft_append_entry(me_, c)) { __log(me_, "AE failure; couldn't append entry"); r.success = 0; goto done; } } r.success = 1; r.current_idx = raft_get_current_idx(me_); r.first_idx = ae->prev_log_idx + 1; done: if (me->cb.send_appendentries_response) me->cb.send_appendentries_response(me_, me->udata, node, &r); return 1; } int raft_recv_requestvote(raft_server_t* me_, int node, msg_requestvote_t* vr) { raft_server_private_t* me = (void*)me_; msg_requestvote_response_t r; if (raft_get_current_term(me_) < vr->term) me->voted_for = -1; if (vr->term < raft_get_current_term(me_) || /* we've already voted */ -1 != me->voted_for || /* we have a more up-to-date log */ vr->last_log_idx < me->current_idx) r.vote_granted = 0; else { raft_vote(me_, node); r.vote_granted = 1; } __log(me_, "node requested vote: %d replying: %s", node, r.vote_granted == 1 ? "granted" : "not granted"); r.term = raft_get_current_term(me_); if (me->cb.send_requestvote_response) me->cb.send_requestvote_response(me_, me->udata, node, &r); return 0; } int raft_votes_is_majority(const int num_nodes, const int nvotes) { int half; if (num_nodes < nvotes) return 0; half = num_nodes / 2; return half + 1 <= nvotes; } int raft_recv_requestvote_response(raft_server_t* me_, int node, msg_requestvote_response_t* r) { raft_server_private_t* me = (void*)me_; __log(me_, "node responded to requestvote: %d status: %s", node, r->vote_granted == 1 ? "granted" : "not granted"); if (raft_is_leader(me_)) return 0; assert(node < me->num_nodes); // if (r->term != raft_get_current_term(me_)) // return 0; if (1 == r->vote_granted) { int votes; me->votes_for_me[node] = 1; votes = raft_get_nvotes_for_me(me_); if (raft_votes_is_majority(me->num_nodes, votes)) raft_become_leader(me_); } return 0; } int raft_send_entry_response(raft_server_t* me_, int node, int etyid, int was_committed) { raft_server_private_t* me = (void*)me_; msg_entry_response_t res; __log(me_, "send entry response to: %d", node); res.id = etyid; res.was_committed = was_committed; if (me->cb.send_entries_response) me->cb.send_entries_response(me_, me->udata, node, &res); return 0; } int raft_recv_entry(raft_server_t* me_, int node, msg_entry_t* e) { raft_server_private_t* me = (void*)me_; raft_entry_t ety; int res, i; __log(me_, "received entry from: %d", node); ety.term = me->current_term; ety.id = e->id; ety.data = e->data; ety.len = e->len; res = raft_append_entry(me_, &ety); raft_send_entry_response(me_, node, e->id, res); for (i = 0; i < me->num_nodes; i++) { if (me->nodeid == i) continue; raft_send_appendentries(me_, i); } return 0; } int raft_send_requestvote(raft_server_t* me_, int node) { raft_server_private_t* me = (void*)me_; msg_requestvote_t rv; __log(me_, "sending requestvote to: %d", node); rv.term = me->current_term; rv.last_log_idx = raft_get_current_idx(me_); if (me->cb.send_requestvote) me->cb.send_requestvote(me_, me->udata, node, &rv); return 1; } int raft_append_entry(raft_server_t* me_, raft_entry_t* c) { raft_server_private_t* me = (void*)me_; if (1 == log_append_entry(me->log, c)) { me->current_idx += 1; return 1; } return 0; } int raft_apply_entry(raft_server_t* me_) { raft_server_private_t* me = (void*)me_; raft_entry_t* e; if (!(e = log_get_from_idx(me->log, me->last_applied_idx + 1))) return 0; __log(me_, "applying log: %d", me->last_applied_idx); me->last_applied_idx++; if (me->commit_idx < me->last_applied_idx) me->commit_idx = me->last_applied_idx; if (me->cb.applylog) me->cb.applylog(me_, me->udata, e->data, e->len); return 1; } void raft_send_appendentries(raft_server_t* me_, int node) { raft_server_private_t* me = (void*)me_; __log(me_, "sending appendentries to: %d", node); if (!(me->cb.send_appendentries)) return; msg_appendentries_t ae; raft_node_t* p = raft_get_node(me_, node); ae.term = me->current_term; ae.leader_id = me->nodeid; ae.prev_log_term = raft_node_get_next_idx(p); // TODO: ae.prev_log_idx = 0; ae.n_entries = 0; if (me->cb.send_appendentries) me->cb.send_appendentries(me_, me->udata, node, &ae); } void raft_send_appendentries_all(raft_server_t* me_) { raft_server_private_t* me = (void*)me_; int i; for (i = 0; i < me->num_nodes; i++) { if (me->nodeid == i) continue; raft_send_appendentries(me_, i); } } void raft_set_configuration(raft_server_t* me_, raft_node_configuration_t* nodes, int my_idx) { raft_server_private_t* me = (void*)me_; int num_nodes; /* TODO: one memory allocation only please */ for (num_nodes = 0; nodes->udata_address; nodes++) { num_nodes++; me->nodes = realloc(me->nodes, sizeof(raft_node_t*) * num_nodes); me->num_nodes = num_nodes; me->nodes[num_nodes - 1] = raft_node_new(nodes->udata_address); } me->votes_for_me = calloc(num_nodes, sizeof(int)); me->nodeid = my_idx; } int raft_get_nvotes_for_me(raft_server_t* me_) { raft_server_private_t* me = (void*)me_; int i, votes; for (i = 0, votes = 0; i < me->num_nodes; i++) { if (me->nodeid == i) continue; if (1 == me->votes_for_me[i]) votes += 1; } if (me->voted_for == me->nodeid) votes += 1; return votes; } void raft_vote(raft_server_t* me_, int node) { raft_server_private_t* me = (void*)me_; me->voted_for = node; } raft_node_t* raft_get_node(raft_server_t *me_, int nodeid) { raft_server_private_t* me = (void*)me_; if (nodeid < 0 || me->num_nodes <= nodeid) return NULL; return me->nodes[nodeid]; } int raft_is_follower(raft_server_t* me_) { return raft_get_state(me_) == RAFT_STATE_FOLLOWER; } int raft_is_leader(raft_server_t* me_) { return raft_get_state(me_) == RAFT_STATE_LEADER; } int raft_is_candidate(raft_server_t* me_) { return raft_get_state(me_) == RAFT_STATE_CANDIDATE; } /*--------------------------------------------------------------79-characters-*/