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

fix(controller): never close() the primary channel's raw fd (bug-1865)

Browse source 
handle_ctrlhello stored the netio endpoint's fd (primary_ctrl_fd = ep->fd)
and ha_on_primary_ctrl_fd_rdhup close()d it in three branches. netio owns
that socket and closes it on endpoint teardown, so by the time the
liveness poll fired the RDHUP the number could be reused -- in a chaotic
soak by the standby's WAL segment: HA's close() shut the WAL, fdatasync
failed 2305 times (EBADF, then EINVAL once a socket took the number),
durable_lsn stuck, and one round's jobsub failed EIO.

handle_ctrlhello now also holds a reference on the endpoint
(primary_ctrl_ep), which keeps its socket open and its fd number ours.
ha_drop_primary_ctrl_conn() replaces the raw closes: shutdown() the
still-owned socket so netio sees the hang-up and tears it down itself,
then release the reference. A bare fd with no endpoint is only forgotten.

Test test_rdhup_never_closes_an_fd_netio_already_released: real
ctrlhello -> netio teardown -> file opened on the freed number -> RDHUP;
the file was closed before, survives now. test_rdhup_closes_and_nulls_fd
updated to the new contract (peer reads EOF, fields cleared). Controller
suite 5/5.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
This commit is contained in:
Rongsong Shen 2026-09-25 14:55:04 +00:00
parent 98c52b7d96
commit 910b43cd37
Signed by untrusted user: rshen
GPG key ID: 80533536B0EB651C
5 changed files with 127 additions and 22 deletions
Download patch file Download diff file Expand all files Collapse all files

27
src/controller/testsuite/test_ha_failover.c
View file

@ -25,6 +25,7 @@
#include <unity/unity.h>
#include "ha.h"
#include "netio.h"
#include "xce/cluster_def.h" /* full cluster_t layout for task-5.2 tests */
#include "myconf.h" /* myconf_t for task-5.2 tests */
@ -91,28 +92,38 @@ static ha_config_t g_fast_cfg = {
.election_priority_step_ms = 500,
};
/* ── Test 1: RDHUP closes fd and sets primary_ctrl_fd = -1 ─────────────── */
/* ── Test 1: RDHUP drops the primary channel and sets primary_ctrl_fd = -1 ──
* bug-1865: the channel is a netio endpoint HA holds a reference on; RDHUP
* shuts its socket down (the peer sees EOF) and releases the reference --
* it never close()s the raw fd, which netio owns. */
static void test_rdhup_closes_and_nulls_fd(void)
{
ha_state_t ha;
ha_state_t ha;
endpoint_t* ep;
char b;
TEST_ASSERT_EQUAL_INT(0, ha_init(&ha, &g_cfg, 1));
/* Set up a socketpair as the "primary_ctrl_fd" */
/* Set up a socketpair as the primary channel, held like handle_ctrlhello */
int sv[2];
TEST_ASSERT_EQUAL_INT(0, socketpair(AF_UNIX, SOCK_STREAM, 0, sv));
ep = endpoint_get();
TEST_ASSERT_NOT_NULL(ep);
ep->fd = sv[0];
endpoint_ref(ep);
ha.primary_ctrl_fd = sv[0];
ha.primary_ctrl_ep = ep;
/* Fire RDHUP with NULL cluster (unit test mode — skips persist/election) */
ha_on_primary_ctrl_fd_rdhup(&ha, NULL);
/* fd must be closed and set to -1 */
/* channel dropped: fields cleared, peer sees the hang-up */
TEST_ASSERT_EQUAL_INT(-1, ha.primary_ctrl_fd);
/* Verify sv[0] is really closed by trying to write to it */
ssize_t n = write(sv[0], "x", 1);
TEST_ASSERT_LESS_THAN_INT(0, (int)n);
TEST_ASSERT_EQUAL_INT(EBADF, errno);
TEST_ASSERT_NULL(ha.primary_ctrl_ep);
TEST_ASSERT_EQUAL_INT(0, (int)read(sv[1], &b, 1));
endpoint_unref(ep); /* netio's own reference: closes sv[0] */
close(sv[1]);
ha_fini(&ha);
printf("test_rdhup_closes_and_nulls_fd: PASS\n");