BUG/MEDIUM: servers: Add a per-thread counter of idle connections.
Add a per-thread counter of idling connections, and use it to determine
how many connections we should kill after the timeout, instead of using
the global counter, or we're likely to just kill most of the connections.
This should be backported to 1.9.
diff --git a/src/backend.c b/src/backend.c
index 3c752a2..d15020a 100644
--- a/src/backend.c
+++ b/src/backend.c
@@ -1344,6 +1344,7 @@
LIST_DEL(&srv_conn->list);
srv_conn->idle_time = 0;
HA_ATOMIC_SUB(&srv->curr_idle_conns, 1);
+ srv->curr_idle_thr[tid]--;
LIST_ADDQ(&srv->idle_conns[tid], &srv_conn->list);
if (LIST_ISEMPTY(&srv->idle_orphan_conns[tid]))
task_unlink_wq(srv->idle_task[tid]);