mirrors/scoutfs
1
0
Fork 0
mirror of https://github.com/versity/scoutfs.git synced 2026-08-15 19:56:36 +00:00
Code Issues Packages Projects Releases Wiki Activity

Only start new quorum election after a receive failure

Browse Source 
It's possible for the quorum worker to be preempted for a long period,
especially on debug kernels. Since we only check for how much time
has passed, it's possible for a clean receive to inadvertently
trigger an election. This can cause the quorum-heartbeat-timeout
test to fail due to observed delays outside of the expected bounds.

Instead, make sure we had a receive failure before comparing timestamps.

Signed-off-by: Chris Kirby <ckirby@versity.com>
This commit is contained in:
Chris Kirby
2025-10-06 12:27:25 -05:00
parent 35bcad91a6
commit 48d849e2f4
1 changed files with 12 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
kmod/src/quorum.c
+12 -1
View File
@@ -726,6 +726,8 @@ static void scoutfs_quorum_worker(struct work_struct *work)
struct quorum_status qst = {0,};
struct hb_recording hbr;
bool record_hb;
bool recv_failed;
bool initializing = true;
int ret;
int err;
@@ -758,6 +760,8 @@ static void scoutfs_quorum_worker(struct work_struct *work)
update_show_status(qinf, &qst);
recv_failed = false;
ret = recv_msg(sb, &msg, qst.timeout);
if (ret < 0) {
if (ret != -ETIMEDOUT && ret != -EAGAIN) {
@@ -765,6 +769,9 @@ static void scoutfs_quorum_worker(struct work_struct *work)
scoutfs_inc_counter(sb, quorum_recv_error);
goto out;
}
recv_failed = true;
msg.type = SCOUTFS_QUORUM_MSG_INVALID;
ret = 0;
}
@@ -822,12 +829,13 @@ static void scoutfs_quorum_worker(struct work_struct *work)
/* followers and candidates start new election on timeout */
if (qst.role != LEADER &&
(initializing || recv_failed) &&
ktime_after(ktime_get(), qst.timeout)) {
/* .. but only if their server has stopped */
if (!scoutfs_server_is_down(sb)) {
qst.timeout = election_timeout();
scoutfs_inc_counter(sb, quorum_candidate_server_stopping);
continue;
goto again;
}
qst.role = CANDIDATE;
@@ -964,6 +972,9 @@ static void scoutfs_quorum_worker(struct work_struct *work)
}
record_hb_delay(sb, qinf, &hbr, record_hb, qst.role);
again:
initializing = false;
}
update_show_status(qinf, &qst);