feat(election-iomgr-refactor): extend handle_request_vote with role-based deny gate — task 3.2

This commit is contained in:
Rongsong Shen 2026-05-30 11:03:35 +00:00
parent e6135ea828
commit 6de4b3ff0e
4 changed files with 286 additions and 1 deletions

View file

@ -824,11 +824,34 @@ int ha_restore_term(ha_state_t *ha, const char *workdir)
return 0;
}
/* ── Vote Handler (Req 2.2–2.6, 9.1) ───────────────────────────────────────
/* ── ha_heartbeat_is_fresh (Req 6.2) ────────────────────────────────────────
*
* Returns 1 if the elapsed time since ha->last_heartbeat is strictly less than
* ha->config.heartbeat_timeout_ms (i.e. the primary is still alive).
* Returns 0 if the heartbeat is stale (timeout has expired).
*
* Uses CLOCK_MONOTONIC to avoid wall-clock skew.
*/
static int ha_heartbeat_is_fresh(ha_state_t *ha)
{
struct timespec now, last;
clock_gettime(CLOCK_MONOTONIC, &now);
pthread_mutex_lock(&ha->lock);
last = ha->last_heartbeat;
pthread_mutex_unlock(&ha->lock);
long ms = (now.tv_sec - last.tv_sec) * 1000L
+ (now.tv_nsec - last.tv_nsec) / 1000000L;
return ms < (long)ha->config.heartbeat_timeout_ms;
}
/* ── Vote Handler (Req 2.2–2.6, 6.1, 6.2, 6.4, 9.1) ────────────────────────
*
* handle_request_vote: process an incoming OP_requestVote RPC.
*
* Decision order (fail-fast):
* 0. role gate (Req 6.1, 6.2):
* - CTRL_PRIMARY or CTRL_PRIMARY_COMMITTING → deny
* - CTRL_STANDBY with fresh heartbeat → deny
* 1. in_catch_up → deny (Req 2.5)
* 2. req term < my term → deny (Req 2.2)
* 3. req term > my term → update term, clear voted_for (Req 2.3)
@ -847,6 +870,17 @@ int handle_request_vote(ha_state_t *ha,
uint64_t my_last_lsn,
int in_catch_up_flag)
{
/* Req 6.1, 6.2: role-based deny gate — checked before any Raft state mutation.
* PRIMARY and PRIMARY_COMMITTING never grant votes (protect the live primary).
* STANDBY with a fresh heartbeat also denies (primary is alive, no new election
* needed). CTRL_INIT and CTRL_ELECTING fall through to Raft evaluation. */
ctrl_role_t my_role = (ctrl_role_t)atomic_load_explicit(&ha->role,
memory_order_acquire);
if (my_role == CTRL_PRIMARY || my_role == CTRL_PRIMARY_COMMITTING)
return 0; /* deny — I am primary (or committing); protect against challengers */
if (my_role == CTRL_STANDBY && ha_heartbeat_is_fresh(ha))
return 0; /* deny — my primary is alive */
/* Req 2.5: catching-up standbys deny all votes */
if (in_catch_up_flag)
return 0;

View file

@ -113,6 +113,7 @@ controller_unit_test_src = [
'testsuite/test_ha_failover.c',
'testsuite/test_ha_vote_send.c',
'testsuite/test_ha_start_election_async.c',
'testsuite/test_ha_vote_gate.c',
'startup_validation.c',
'worker.c',
'persist.c',

View file

@ -45,6 +45,7 @@ extern void test_ha_ctrlhello(void);
extern void test_ha_failover(void);
extern void test_ha_vote_send(void);
extern void test_ha_start_election_async(void);
extern void test_ha_vote_gate(void);
void setUp(void) { return; }
@ -159,6 +160,7 @@ int main(int argc, char* argv[])
TEST(test_ha_failover);
TEST(test_ha_vote_send);
TEST(test_ha_start_election_async);
TEST(test_ha_vote_gate);
return UNITY_END();
}

View file

@ -0,0 +1,248 @@
/* test_ha_vote_gate.c — Unity tests for Task 3.2: role-based vote-deny gate
* in handle_request_vote().
*
* Tests verify (Req 6.1, 6.2, 6.4):
* Req 6.1: CTRL_PRIMARY → deny all vote requests (return 0)
* Req 6.1: CTRL_PRIMARY_COMMITTING → deny all vote requests (return 0)
* Req 6.2: CTRL_STANDBY + fresh heartbeat → deny (return 0)
* Req 6.4: CTRL_STANDBY + stale heartbeat → evaluate Raft gates (not denied by role gate)
* Req 6.4: CTRL_ELECTING → evaluate Raft gates (not denied by role gate)
*
* Implementation gated on:
* - ha_heartbeat_is_fresh(ha): returns 1 if elapsed < heartbeat_timeout_ms
* - handle_request_vote() prepends role-based deny before existing Raft gates
*/
#define _GNU_SOURCE
#include <stdatomic.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <time.h>
#include <unistd.h>
#include <unity/unity.h>
#include "controller.capnp.h"
#include "ha.h"
/* Functions under test — declared here to match their signatures in ha.c */
int handle_request_vote(ha_state_t *ha,
uint64_t term,
const char *candidate_id,
uint64_t config_tag,
uint64_t last_lsn,
uint64_t my_config_tag,
uint64_t my_last_lsn,
int in_catch_up_flag);
/* ── Shared config for vote gate tests ──────────────────────────────────── */
static ha_config_t g_gate_cfg = {
.heartbeat_interval_ms = 3000,
.heartbeat_timeout_ms = 10000, /* 10 s — used for ha_heartbeat_is_fresh */
.election_timeout_ms = 2000,
.probe_connect_timeout_ms = 500,
.max_promotion_retries = 5,
.election_jitter_max_ms = 0,
.election_priority_step_ms = 500,
};
/* ── Helper: set ha->last_heartbeat to "now - offset_ms milliseconds ago" ── */
static void set_last_heartbeat_age(ha_state_t *ha, long offset_ms)
{
struct timespec now;
clock_gettime(CLOCK_MONOTONIC, &now);
long sec = offset_ms / 1000;
long nsec = (offset_ms % 1000) * 1000000L;
ha->last_heartbeat.tv_sec = now.tv_sec - sec;
ha->last_heartbeat.tv_nsec = now.tv_nsec - nsec;
/* Carry / borrow normalisation */
if (ha->last_heartbeat.tv_nsec < 0) {
ha->last_heartbeat.tv_sec -= 1;
ha->last_heartbeat.tv_nsec += 1000000000L;
}
}
/* ── Req 6.1: CTRL_PRIMARY → deny regardless of candidate freshness ────── */
static void test_primary_role_denies_vote_request(void)
{
/* When the receiving node is CTRL_PRIMARY it must deny all vote requests
* without evaluating term, config_tag, or LSN. The candidate is perfectly
* fresh (same config_tag, same LSN, same term) — still deny. */
ha_state_t ha;
ha_init(&ha, &g_gate_cfg, 0);
atomic_store_explicit(&ha.role, (int32_t)CTRL_PRIMARY, memory_order_release);
atomic_store_explicit(&ha.term, 5ULL, memory_order_release);
ha.voted_for_term = 0;
int granted = handle_request_vote(&ha,
5ULL, /* same term — would pass term check */
"ctrl-b",
100ULL, /* same config_tag as mine */
50ULL, /* same LSN as mine */
100ULL, /* my config_tag */
50ULL, /* my last_lsn */
0 /* not in catch_up */);
TEST_ASSERT_EQUAL_INT_MESSAGE(0, granted,
"CTRL_PRIMARY must deny all incoming vote requests (Req 6.1)");
ha_fini(&ha);
printf("test_primary_role_denies_vote_request: PASS\n");
}
/* ── Req 6.1: CTRL_PRIMARY_COMMITTING → deny ────────────────────────────── */
static void test_primary_committing_role_denies_vote_request(void)
{
/* When the receiving node is CTRL_PRIMARY_COMMITTING (waiting for quorum
* replication ack before opening traffic), it must also deny all vote
* requests. A challenger must not displace a committing primary. */
ha_state_t ha;
ha_init(&ha, &g_gate_cfg, 0);
atomic_store_explicit(&ha.role, (int32_t)CTRL_PRIMARY_COMMITTING,
memory_order_release);
atomic_store_explicit(&ha.term, 3ULL, memory_order_release);
ha.voted_for_term = 0;
int granted = handle_request_vote(&ha,
3ULL, /* same term */
"ctrl-c",
200ULL, /* candidate is fresh */
100ULL,
100ULL, /* my config_tag */
50ULL,
0);
TEST_ASSERT_EQUAL_INT_MESSAGE(0, granted,
"CTRL_PRIMARY_COMMITTING must deny all vote requests (Req 6.1)");
ha_fini(&ha);
printf("test_primary_committing_role_denies_vote_request: PASS\n");
}
/* ── Req 6.2: CTRL_STANDBY + fresh heartbeat → deny ────────────────────── */
static void test_standby_fresh_heartbeat_denies_vote_request(void)
{
/* When the receiving node is CTRL_STANDBY and received a heartbeat recently
* (within heartbeat_timeout_ms = 10 s), ha_heartbeat_is_fresh() returns 1
* and the function must deny the vote — my primary is alive. */
ha_state_t ha;
ha_init(&ha, &g_gate_cfg, 0);
atomic_store_explicit(&ha.role, (int32_t)CTRL_STANDBY, memory_order_release);
atomic_store_explicit(&ha.term, 2ULL, memory_order_release);
ha.voted_for_term = 0;
/* last_heartbeat = 500 ms ago (< 10 000 ms timeout → fresh) */
set_last_heartbeat_age(&ha, 500L);
int granted = handle_request_vote(&ha,
2ULL,
"ctrl-b",
100ULL,
50ULL,
100ULL,
50ULL,
0);
TEST_ASSERT_EQUAL_INT_MESSAGE(0, granted,
"CTRL_STANDBY with fresh heartbeat must deny vote requests (Req 6.2)");
ha_fini(&ha);
printf("test_standby_fresh_heartbeat_denies_vote_request: PASS\n");
}
/* ── Req 6.4: CTRL_STANDBY + stale heartbeat → evaluate Raft gates ──────── */
static void test_standby_stale_heartbeat_evaluates_vote(void)
{
/* When the receiving node is CTRL_STANDBY but the heartbeat is older than
* heartbeat_timeout_ms (10 s), ha_heartbeat_is_fresh() returns 0.
* The role gate must NOT deny — fall through to evaluate Raft gates.
*
* We set up a candidate with the same config_tag and LSN (freshness tie →
* grant) and no prior vote — the Raft logic should grant the vote. */
ha_state_t ha;
ha_init(&ha, &g_gate_cfg, 0);
atomic_store_explicit(&ha.role, (int32_t)CTRL_STANDBY, memory_order_release);
atomic_store_explicit(&ha.term, 1ULL, memory_order_release);
ha.voted_for_term = 0;
/* last_heartbeat = 15 000 ms ago (> 10 000 ms timeout → stale) */
set_last_heartbeat_age(&ha, 15000L);
int granted = handle_request_vote(&ha,
1ULL,
"ctrl-a",
100ULL, /* same config_tag — freshness tie */
50ULL, /* same LSN — freshness tie */
100ULL,
50ULL,
0);
/* With stale heartbeat the role gate does not deny; Raft gates evaluate
* and should grant (equal freshness, no prior vote). */
TEST_ASSERT_EQUAL_INT_MESSAGE(1, granted,
"CTRL_STANDBY with stale heartbeat must evaluate Raft gates (Req 6.4)");
ha_fini(&ha);
printf("test_standby_stale_heartbeat_evaluates_vote: PASS\n");
}
/* ── Req 6.4: CTRL_ELECTING → evaluate Raft gates ───────────────────────── */
static void test_electing_role_evaluates_vote(void)
{
/* When the receiving node is CTRL_ELECTING (it is itself campaigning),
* the role gate must not deny — fall through to Raft term/freshness gates.
*
* A valid candidate (same term, fresh config_tag, no prior vote) must be
* granted, exercising the full Raft evaluation path. */
ha_state_t ha;
ha_init(&ha, &g_gate_cfg, 0);
atomic_store_explicit(&ha.role, (int32_t)CTRL_ELECTING, memory_order_release);
atomic_store_explicit(&ha.term, 4ULL, memory_order_release);
ha.voted_for_term = 0;
int granted = handle_request_vote(&ha,
4ULL,
"ctrl-c",
300ULL, /* candidate fresher (higher tag) */
200ULL,
100ULL, /* my config_tag (lower) */
50ULL,
0);
/* Role gate does not fire; Raft gates evaluate; candidate is fresh → grant */
TEST_ASSERT_EQUAL_INT_MESSAGE(1, granted,
"CTRL_ELECTING must evaluate Raft gates, not deny via role gate (Req 6.4)");
ha_fini(&ha);
printf("test_electing_role_evaluates_vote: PASS\n");
}
/* ── Test suite entry point ───────────────────────────────────────────────── */
void test_ha_vote_gate(void)
{
/* Req 6.1 — PRIMARY denies */
RUN_TEST(test_primary_role_denies_vote_request);
/* Req 6.1 — PRIMARY_COMMITTING denies */
RUN_TEST(test_primary_committing_role_denies_vote_request);
/* Req 6.2 — STANDBY + fresh heartbeat denies */
RUN_TEST(test_standby_fresh_heartbeat_denies_vote_request);
/* Req 6.4 — STANDBY + stale heartbeat evaluates */
RUN_TEST(test_standby_stale_heartbeat_evaluates_vote);
/* Req 6.4 — ELECTING evaluates */
RUN_TEST(test_electing_role_evaluates_vote);
}