protocol/server: make listen backlog value as configurable

problem:

When we call listen from protocol/server, we are giving a
hard coded valie of 10 if it is not manually given.
With multiplexing, especially when glusterd restarts all
clients may try to connect to the server at a time.
Which will result in overflowing the queue, and kernel
will complain about the errors.

Solution:

This patch will introduce a volume set command to make backlog
value as a configurable. This patch also changes the default
values for backlog from 10 to 128. This changes is only applicable
for sockets listening from protocol.

Example:

gluster volume set <volname> transport.listen-backlog 1024

Note: 1 Brick has to be restarted to get this value in effect
      2 This changes won't be reflected in glusterd, or other
        xlators which calls listen. If you need, you have to
        add this option to the volfile.

Change-Id: I0c5a2bbf28b5db612f9979e7560e05dd82b41477
BUG: 1456405
Signed-off-by: Mohammed Rafi KC <rkavunga@redhat.com>
Reviewed-on: https://review.gluster.org/17411
Smoke: Gluster Build System <jenkins@build.gluster.org>
NetBSD-regression: NetBSD Build System <jenkins@build.gluster.org>
CentOS-regression: Gluster Build System <jenkins@build.gluster.org>
Reviewed-by: Raghavendra G <rgowdapp@redhat.com>
Reviewed-by: Raghavendra Talur <rtalur@redhat.com>
Reviewed-by: Atin Mukherjee <amukherj@redhat.com>
Reviewed-by: Niels de Vos <ndevos@redhat.com>
Reviewed-by: Jeff Darcy <jeff@pl.atyp.us>

3 files changed, 47 insertions, 11 deletions

diff --git a/rpc/rpc-transport/rdma/src/rdma.c b/rpc/rpc-transport/rdma/src/rdma.c
index 8d9e6474f3c..01f96c21b5c 100644
--- a/rpc/rpc-transport/rdma/src/rdma.c
+++ b/rpc/rpc-transport/rdma/src/rdma.c
@@ -4499,6 +4499,12 @@ gf_rdma_options_init (rpc_transport_t *this)
         options->attr_retry_cnt = GF_RDMA_RETRY_CNT;
         options->attr_rnr_retry = GF_RDMA_RNR_RETRY;
 
+        temp = dict_get (this->options, "transport.listen-backlog");
+        if (temp)
+                options->backlog = data_to_uint32 (temp);
+        else
+                options->backlog = GLUSTERFS_SOCKET_LISTEN_BACKLOG;
+
         temp = dict_get (this->options,
                          "transport.rdma.work-request-send-count");
         if (temp)
@@ -4635,6 +4641,7 @@ gf_rdma_init (rpc_transport_t *this)
         priv->peer.recv_count = options->recv_count;
         priv->peer.send_size = options->send_size;
         priv->peer.recv_size = options->recv_size;
+        priv->backlog = options->backlog;
 
         priv->peer.trans = this;
         INIT_LIST_HEAD (&priv->peer.ioq);
@@ -4848,7 +4855,8 @@ gf_rdma_listen (rpc_transport_t *this)
                 goto err;
         }
 
-        ret = rdma_listen (peer->cm_id, 10);
+        ret = rdma_listen (peer->cm_id, priv->backlog);
+
         if (ret != 0) {
                 gf_msg (this->name, GF_LOG_WARNING, errno,
                         RDMA_MSG_LISTEN_FAILED,
@@ -4919,6 +4927,28 @@ init (rpc_transport_t *this)
         return 0;
 }
 
+int
+reconfigure (rpc_transport_t *this, dict_t *options)
+{
+        gf_rdma_private_t *priv          = NULL;
+        uint32_t          backlog        = 0;
+        int               ret            = -1;
+
+        GF_VALIDATE_OR_GOTO ("rdma", this, out);
+        GF_VALIDATE_OR_GOTO ("rdma", this->private, out);
+
+        priv = this->private;
+
+        if (dict_get_uint32 (options, "transport.listen-backlog",
+                             &backlog) == 0) {
+                priv->backlog = backlog;
+                gf_log (this->name, GF_LOG_DEBUG, "Reconfigued "
+                        "transport.listen-backlog=%d", priv->backlog);
+        }
+        ret = 0;
+out:
+        return ret;
+}
 void
 fini (struct rpc_transport *this)
 {
diff --git a/rpc/rpc-transport/rdma/src/rdma.h b/rpc/rpc-transport/rdma/src/rdma.h
index 449861f075f..6bcf6bc6ef2 100644
--- a/rpc/rpc-transport/rdma/src/rdma.h
+++ b/rpc/rpc-transport/rdma/src/rdma.h
@@ -144,6 +144,7 @@ struct __gf_rdma_options {
 	uint8_t  attr_timeout;
 	uint8_t  attr_retry_cnt;
 	uint8_t  attr_rnr_retry;
+        uint32_t backlog;
 };
 typedef struct __gf_rdma_options gf_rdma_options_t;
 
@@ -383,6 +384,7 @@ struct __gf_rdma_private {
         pthread_mutex_t             recv_mutex;
         pthread_cond_t              recv_cond;
         gf_rdma_transport_entity_t  entity;
+        uint32_t                    backlog;
 };
 typedef struct __gf_rdma_private    gf_rdma_private_t;
 
diff --git a/rpc/rpc-transport/socket/src/socket.c b/rpc/rpc-transport/socket/src/socket.c
index f5062fb026c..dd7f5d3b77e 100644
--- a/rpc/rpc-transport/socket/src/socket.c
+++ b/rpc/rpc-transport/socket/src/socket.c
@@ -3536,10 +3536,7 @@ socket_listen (rpc_transport_t *this)
                         goto unlock;
                 }
 
-                if (priv->backlog)
-                        ret = listen (priv->sock, priv->backlog);
-                else
-                        ret = listen (priv->sock, 10);
+                ret = listen (priv->sock, priv->backlog);
 
                 if (ret == -1) {
                         gf_log (this->name, GF_LOG_ERROR,
@@ -3853,6 +3850,7 @@ reconfigure (rpc_transport_t *this, dict_t *options)
         gf_boolean_t      tmp_bool      = _gf_false;
         char             *optstr        = NULL;
         int               ret           = 0;
+        uint32_t          backlog       = 0;
         uint64_t          windowsize    = 0;
         uint32_t          timeout       = 0;
         int               keepaliveidle  = GF_KEEPALIVE_TIME;
@@ -3892,6 +3890,13 @@ reconfigure (rpc_transport_t *this, dict_t *options)
         gf_log (this->name, GF_LOG_DEBUG, "Reconfigued "
                 "transport.tcp-user-timeout=%d", priv->timeout);
 
+        if (dict_get_uint32 (options, "transport.listen-backlog",
+                             &backlog) == 0) {
+                priv->backlog = backlog;
+                gf_log (this->name, GF_LOG_DEBUG, "Reconfigued "
+                        "transport.listen-backlog=%d", priv->backlog);
+        }
+
         if (dict_get_int32 (options, "transport.socket.keepalive-time",
                             &(priv->keepaliveidle)) != 0)
                 priv->keepaliveidle = keepaliveidle;
@@ -4190,10 +4195,12 @@ socket_init (rpc_transport_t *this)
                 "transport.keepalivecnt=%d", keepalivecnt);
 
         if (dict_get_uint32 (this->options,
-                             "transport.socket.listen-backlog",
-                             &backlog) == 0) {
-                priv->backlog = backlog;
+                             "transport.listen-backlog",
+                             &backlog) != 0) {
+
+                backlog = GLUSTERFS_SOCKET_LISTEN_BACKLOG;
         }
+        priv->backlog = backlog;
 
         optstr = NULL;
 
@@ -4612,9 +4619,6 @@ struct volume_options options[] = {
           .type  = GF_OPTION_TYPE_INT,
           .default_value = "9"
         },
-        { .key   = {"transport.socket.listen-backlog"},
-          .type  = GF_OPTION_TYPE_INT
-        },
         { .key   = {"transport.socket.read-fail-log"},
           .type  = GF_OPTION_TYPE_BOOL
         },