BUG/MINOR: checks: properly handle wrapping time in __health_adjust()

There's an issue when a server state changes, we use an integer comparison
to decide whether or not to reschedule a test instead of using a wrapping
timer comparison. This will cause some health-checks not to be immediately
triggered half of the time, and some unneeded calls to task_queue() to be
performed in other cases.

This bug has always been there as it was introduced with the commit that
added the feature, 97f07b832 ("[MEDIUM] Decrease server health based on
http responses / events, version 3"). This may be backported everywhere.

(cherry picked from commit 64ba5ebadcd5d98e00989d08dfaa3c94c15196c9)
Signed-off-by: Christopher Faulet <cfaulet@haproxy.com>

diff --git a/src/check.c b/src/check.c
index b95b4ae..37b7d5e 100644 (file)
--- a/src/check.c
+++ b/src/check.c
@@ -473,7 +473,7 @@ void __health_adjust(struct server *s, short status)
 
        if (s->check.fastinter) {
                expire = tick_add(now_ms, MS_TO_TICKS(s->check.fastinter));
-               if (s->check.task->expire > expire) {
+               if (tick_is_lt(expire, s->check.task->expire)) {
                        s->check.task->expire = expire;
                        /* requeue check task with new expire */
                        task_queue(s->check.task);