From 2f50486f431b37f4a16541911f50f2042049b0e6 Mon Sep 17 00:00:00 2001
From: Amar Tumballi <amarts@redhat.com>
Date: Tue, 21 May 2019 01:37:47 +0000
Subject: Revert "rpc: implement reconnect back-off strategy"

This reverts commit 59841f7e1ff0511b04884015441a181a56d07bea.

This revert is done as a 'possible' fix for frequent regression
failures, which are random in nature too (ie, different tests fails
in different runs).

Why exactly this patch? Because this patch seemed like most probable
candidate which got merged in last 15days, and after which regressions
are failing more often.

Updates: bz#1711827
Change-Id: I35333162fcd4064f9609525ca93c666053c6d959
---
 rpc/rpc-lib/src/rpc-clnt.c | 33 ++++++++++++++++-----------------
 rpc/rpc-lib/src/rpc-clnt.h |  1 -
 2 files changed, 16 insertions(+), 18 deletions(-)

diff --git a/rpc/rpc-lib/src/rpc-clnt.c b/rpc/rpc-lib/src/rpc-clnt.c
index c1945dfb6ec..8ef05378351 100644
--- a/rpc/rpc-lib/src/rpc-clnt.c
+++ b/rpc/rpc-lib/src/rpc-clnt.c
@@ -392,16 +392,8 @@ rpc_clnt_reconnect(void *conn_ptr)
         conn->reconnect = 0;
 
         if ((conn->connected == 0) && !clnt->disabled) {
-            if (conn->reconnect_delay.tv_sec < 3) {
-                conn->reconnect_delay.tv_sec *= 2;
-                int64_t ns = conn->reconnect_delay.tv_nsec * 2;
-                if (ns >= 1000000000ULL) {
-                    conn->reconnect_delay.tv_sec++;
-                    ns -= 1000000000ULL;
-                }
-                conn->reconnect_delay.tv_nsec = ns;
-            }
-            ts = conn->reconnect_delay;
+            ts.tv_sec = 3;
+            ts.tv_nsec = 0;
 
             gf_log(conn->name, GF_LOG_TRACE, "attempting reconnect");
             (void)rpc_transport_connect(trans, conn->config.remote_port);
@@ -846,11 +838,9 @@ rpc_clnt_handle_disconnect(struct rpc_clnt *clnt, rpc_clnt_connection_t *conn)
 
     pthread_mutex_lock(&conn->lock);
     {
-        conn->reconnect_delay.tv_sec = 0;
-        conn->reconnect_delay.tv_nsec = 100000000;
-
         if (!conn->rpc_clnt->disabled && (conn->reconnect == NULL)) {
-            ts = conn->reconnect_delay;
+            ts.tv_sec = 3;
+            ts.tv_nsec = 0;
 
             rpc_clnt_ref(clnt);
             conn->reconnect = gf_timer_call_after(clnt->ctx, ts,
@@ -1170,8 +1160,6 @@ rpc_clnt_start(struct rpc_clnt *rpc)
      * rpc_clnt_reconnect fire event.
      */
     rpc_clnt_ref(rpc);
-    conn->reconnect_delay.tv_sec = 0;
-    conn->reconnect_delay.tv_nsec = 50000000;
     rpc_clnt_reconnect(conn);
 
     return 0;
@@ -1189,7 +1177,18 @@ rpc_clnt_cleanup_and_start(struct rpc_clnt *rpc)
 
     rpc_clnt_connection_cleanup(conn);
 
-    return rpc_clnt_start(rpc);
+    pthread_mutex_lock(&conn->lock);
+    {
+        rpc->disabled = 0;
+    }
+    pthread_mutex_unlock(&conn->lock);
+    /* Corresponding unref will be either on successful timer cancel or last
+     * rpc_clnt_reconnect fire event.
+     */
+    rpc_clnt_ref(rpc);
+    rpc_clnt_reconnect(conn);
+
+    return 0;
 }
 
 int
diff --git a/rpc/rpc-lib/src/rpc-clnt.h b/rpc/rpc-lib/src/rpc-clnt.h
index 2c252d5ff86..b46feed50c8 100644
--- a/rpc/rpc-lib/src/rpc-clnt.h
+++ b/rpc/rpc-lib/src/rpc-clnt.h
@@ -136,7 +136,6 @@ struct rpc_clnt_connection {
     struct saved_frames *saved_frames;
     struct timespec last_sent;
     struct timespec last_received;
-    struct timespec reconnect_delay;
     uint64_t pingcnt;
     uint64_t msgcnt;
     uint64_t cleanup_gen;
-- 
cgit