Upstream keepalive: clean read delayed flag in stored connections.

If a connection with the read delayed flag set was stored in the keepalive
cache, and after picking it from the cache a read timer was set on that
connection, this timer was considered a delay timer rather than a socket read
event timer as expected.  The latter timeout is usually much longer than the
former, which caused a significant delay in request processing.

The issue manifested itself with proxy_limit_rate and upstream keepalive
enabled and exists since 973ee2276300 (1.7.7) when proxy_limit_rate was
introduced.

diff --git a/src/http/modules/ngx_http_upstream_keepalive_module.c b/src/http/modules/ngx_http_upstream_keepalive_module.c
index 0048e6bce2b51304f547298af685183bca474e14..90a226d91feab7de609a822ce59c2d3e02a5f181 100644 (file)
--- a/src/http/modules/ngx_http_upstream_keepalive_module.c
+++ b/src/http/modules/ngx_http_upstream_keepalive_module.c
@@ -340,6 +340,7 @@ ngx_http_upstream_free_keepalive_peer(ngx_peer_connection_t *pc, void *data,
     pc->connection = NULL;
 
     if (c->read->timer_set) {
+        c->read->delayed = 0;
         ngx_del_timer(c->read);
     }
     if (c->write->timer_set) {