lei.chen/myxce
1
0
Fork 1
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

fix(ha-prevote): make PreVote denial log-reason match the eligibility gate order exactly

Browse source 
The cluster.c denial-reason elimination ordered catch_up before fresh_heartbeat,
disagreeing with raft_cb_vote_eligible (which checks fresh-heartbeat-on-STANDBY
before catch-up). In the rare STANDBY catch-up + fresh-heartbeat overlap the log
could report catch_up while the gate denied on fresh_heartbeat. Expose the pure
read predicate ha_heartbeat_is_fresh and mirror the gate's exact order so the
logged reason can no longer disagree with the actual denial. Decision unchanged;
log-string fidelity only. prevote/vote-gate/failover-prevote suites green.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This commit is contained in:
Rongsong Shen 2026-06-16 07:48:50 +00:00
parent c45f80ab2b
commit 47f733891b
3 changed files with 16 additions and 10 deletions
Download patch file Download diff file Expand all files Collapse all files

17
src/controller/cluster.c
View file

@ -508,12 +508,12 @@ static int myreq_handler(vtask_t* task)
/* Distinguish the denial reason for operator diagnosis /* Distinguish the denial reason for operator diagnosis
* (Req 7.2). raft_on_pre_vote_request denies on (a) stale term * (Req 7.2). raft_on_pre_vote_request denies on (a) stale term
* — cand_term < raft.term — or (b) the shared eligibility gate * — cand_term < raft.term — or (b) the shared eligibility gate
* (raft_cb_vote_eligible): PRIMARY-deny, fresh-heartbeat * (raft_cb_vote_eligible). Mirror that gate's exact predicate
* stickiness, or catch-up. We re-derive which fired WITHOUT * ORDER so the logged reason cannot disagree with the actual
* calling the file-static heartbeat-fresh predicate: when the * denial: stale-term, then PRIMARY, then fresh-heartbeat on a
* role is STANDBY and the node is neither catching up nor at a * STANDBY, then catch-up. ha_heartbeat_is_fresh is a pure read
* stale term, the only remaining eligibility denial is fresh * exposed for this purpose, so the fresh-heartbeat-vs-catch-up
* heartbeat (a stale heartbeat would have passed the gate). */ * overlap is reported exactly as the gate decided it. */
uint64_t my_raft_term = c->ha.raft.term; uint64_t my_raft_term = c->ha.raft.term;
if (cand_term < my_raft_term) { if (cand_term < my_raft_term) {
deny_reason = "stale_term"; deny_reason = "stale_term";
@ -522,11 +522,12 @@ static int myreq_handler(vtask_t* task)
&c->ha.role, memory_order_acquire); &c->ha.role, memory_order_acquire);
if (role == CTRL_PRIMARY) if (role == CTRL_PRIMARY)
deny_reason = "primary"; deny_reason = "primary";
else if (role == CTRL_STANDBY &&
ha_heartbeat_is_fresh(&c->ha))
deny_reason = "fresh_heartbeat";
else if (atomic_load_explicit(&c->ha.in_catch_up, else if (atomic_load_explicit(&c->ha.in_catch_up,
memory_order_acquire)) memory_order_acquire))
deny_reason = "catch_up"; deny_reason = "catch_up";
else if (role == CTRL_STANDBY)
deny_reason = "fresh_heartbeat";
else else
deny_reason = "eligibility"; deny_reason = "eligibility";
} }

4
src/controller/ha.c
View file

@ -87,7 +87,7 @@ static int vote_ep_err_handler(void *ep_raw);
* PRIMARY registers the persist hooks and arms the own-term gate. */ * PRIMARY registers the persist hooks and arms the own-term gate. */
static void ha_register_persist_hooks(cluster_t *c); static void ha_register_persist_hooks(cluster_t *c);
static void ha_arm_own_term_gate(cluster_t *c); static void ha_arm_own_term_gate(cluster_t *c);
static int ha_heartbeat_is_fresh(ha_state_t *ha); int ha_heartbeat_is_fresh(ha_state_t *ha); /* exposed for PreVote denial-reason logging (cluster.c) */
static int primary_heartbeat_fn(void *arg); static int primary_heartbeat_fn(void *arg);
static void on_standby_connected(endpoint_t *ep); static void on_standby_connected(endpoint_t *ep);
@ -2093,7 +2093,7 @@ int ha_restore_term(ha_state_t *ha, const char *workdir)
* Uses CLOCK_MONOTONIC seconds to avoid wall-clock skew; whole-second * Uses CLOCK_MONOTONIC seconds to avoid wall-clock skew; whole-second
* granularity is sufficient for the 10 s default timeout. * granularity is sufficient for the 10 s default timeout.
*/ */
static int ha_heartbeat_is_fresh(ha_state_t *ha) int ha_heartbeat_is_fresh(ha_state_t *ha)
{ {
time_t last = ha->last_activity; /* time_t load is atomic on LP64 */ time_t last = ha->last_activity; /* time_t load is atomic on LP64 */
if (last == 0) if (last == 0)

5
src/controller/ha.h
View file

@ -763,6 +763,11 @@ void ha_discover_primary(ha_state_t *ha, cluster_t *c);
/* ha_on_heartbeat_received: reset the heartbeat deadline (Req 6.3). */ /* ha_on_heartbeat_received: reset the heartbeat deadline (Req 6.3). */
void ha_on_heartbeat_received(ha_state_t *ha); void ha_on_heartbeat_received(ha_state_t *ha);
/* ha_heartbeat_is_fresh: 1 if a primary message arrived within
* heartbeat_timeout_ms (leader-stickiness predicate). Exposed for PreVote
* denial-reason logging in cluster.c; pure read, no state change. */
int ha_heartbeat_is_fresh(ha_state_t *ha);
/* ha_disarm_heartbeat_timeout_vtask: stop the heartbeat watchdog thread. */ /* ha_disarm_heartbeat_timeout_vtask: stop the heartbeat watchdog thread. */
void ha_disarm_heartbeat_timeout_vtask(ha_state_t *ha); void ha_disarm_heartbeat_timeout_vtask(ha_state_t *ha);