transport/socket: configuring tcp window-size

Till now, send and recieve buffer window sizes for sockets
were set to a default glusterfs-specific value.
Linux's default window sizes have been found to be better
w.r.t performance, and hence, no more setting it to any
default value.

However, if one wishes, there's the new configuration option:
   network.tcp-window-size <sane_size>
which takes a size value (int or human readable) and will set
the window size of sockets for both clients and servers.
Nfs clients will also be updated with the same.

Change-Id: I841479bbaea791b01086c42f58401ed297ff16ea
BUG: 795635
Signed-off-by: Rajesh Amaravathi <rajesh@redhat.com>
Reviewed-on: http://review.gluster.com/2821
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Amar Tumballi <amarts@redhat.com>
Reviewed-by: Jeff Darcy <jdarcy@redhat.com>
Reviewed-by: Vijay Bellur <vijay@gluster.com>

7 files changed, 69 insertions, 52 deletions

diff --git a/xlators/mgmt/glusterd/src/glusterd-op-sm.c b/xlators/mgmt/glusterd/src/glusterd-op-sm.c
index da1299de066..3c1baa7e62e 100644
--- a/xlators/mgmt/glusterd/src/glusterd-op-sm.c
+++ b/xlators/mgmt/glusterd/src/glusterd-op-sm.c
@@ -252,7 +252,7 @@ glusterd_op_stage_set_volume (dict_t *dict, char **op_errstr)
         int                                      ret           = 0;
         char                                    *volname       = NULL;
         int                                      exists        = 0;
-        char                                    *key               = NULL;
+        char                                    *key           = NULL;
         char                                    *key_fixed     = NULL;
         char                                    *value         = NULL;
         char                                     str[100]      = {0, };
@@ -279,7 +279,8 @@ glusterd_op_stage_set_volume (dict_t *dict, char **op_errstr)
         ret = dict_get_int32 (dict, "count", &dict_count);
 
         if (ret) {
-                gf_log ("", GF_LOG_ERROR, "Count(dict),not set in Volume-Set");
+                gf_log (this->name, GF_LOG_ERROR,
+                        "Count(dict),not set in Volume-Set");
                 goto out;
         }
 
@@ -297,14 +298,14 @@ glusterd_op_stage_set_volume (dict_t *dict, char **op_errstr)
                         goto out;
 #else
                         ret  = -1;
-                        gf_log ("", GF_LOG_ERROR, "libxml not present in the"
-                                   "system");
+                        gf_log (this->name, GF_LOG_ERROR,
+                                "libxml not present in the system");
                         *op_errstr = gf_strdup ("Error: xml libraries not "
                                                 "present to produce xml-output");
                         goto out;
 #endif
                 }
-                gf_log ("", GF_LOG_ERROR, "No options received ");
+                gf_log (this->name, GF_LOG_ERROR, "No options received ");
                 *op_errstr = gf_strdup ("Options not specified");
                 ret = -1;
                 goto out;
@@ -312,7 +313,7 @@ glusterd_op_stage_set_volume (dict_t *dict, char **op_errstr)
 
         ret = dict_get_str (dict, "volname", &volname);
         if (ret) {
-                gf_log ("", GF_LOG_ERROR, "Unable to get volume name");
+                gf_log (this->name, GF_LOG_ERROR, "Unable to get volume name");
                 goto out;
         }
 
@@ -320,7 +321,7 @@ glusterd_op_stage_set_volume (dict_t *dict, char **op_errstr)
         if (!exists) {
                 snprintf (errstr, sizeof (errstr), "Volume %s does not exist",
                           volname);
-                gf_log ("", GF_LOG_ERROR, "%s", errstr);
+                gf_log (this->name, GF_LOG_ERROR, "%s", errstr);
                 *op_errstr = gf_strdup (errstr);
                 ret = -1;
                 goto out;
@@ -328,7 +329,8 @@ glusterd_op_stage_set_volume (dict_t *dict, char **op_errstr)
 
         ret = glusterd_volinfo_find (volname, &volinfo);
         if (ret) {
-                gf_log ("", GF_LOG_ERROR, "Unable to allocate memory");
+                gf_log (this->name, GF_LOG_ERROR,
+                        "Unable to allocate memory");
                 goto out;
         }
 
@@ -347,7 +349,8 @@ glusterd_op_stage_set_volume (dict_t *dict, char **op_errstr)
                         goto out;
                 }
                 if (!exists) {
-                        gf_log ("", GF_LOG_ERROR, "Option with name: %s "
+                        gf_log (this->name, GF_LOG_ERROR,
+                                "Option with name: %s "
                                 "does not exist", key);
                         ret = snprintf (errstr, 2048,
                                        "option : %s does not exist",
@@ -364,7 +367,7 @@ glusterd_op_stage_set_volume (dict_t *dict, char **op_errstr)
                 ret = dict_get_str (dict, str, &value);
 
                 if (ret) {
-                        gf_log ("", GF_LOG_ERROR,
+                        gf_log (this->name, GF_LOG_ERROR,
                                 "invalid key,value pair in 'volume set'");
                         ret = -1;
                         goto out;
@@ -380,7 +383,7 @@ glusterd_op_stage_set_volume (dict_t *dict, char **op_errstr)
                 ret = dict_set_str (val_dict, key, value);
 
                 if (ret) {
-                        gf_log ("", GF_LOG_ERROR,
+                        gf_log (this->name, GF_LOG_ERROR,
                                 "Unable to set the options in 'volume set'");
                         ret = -1;
                         goto out;
@@ -399,7 +402,7 @@ glusterd_op_stage_set_volume (dict_t *dict, char **op_errstr)
                 }
 
                 if (ret) {
-                        gf_log ("glusterd", GF_LOG_DEBUG, "Could not create temp "
+                        gf_log (this->name, GF_LOG_DEBUG, "Could not create temp "
                                 "volfile, some option failed: %s", *op_errstr);
                         goto out;
                 }
@@ -424,11 +427,11 @@ out:
         if (ret) {
                 if (!(*op_errstr)) {
                         *op_errstr = gf_strdup ("Error, Validation Failed");
-                        gf_log ("glsuterd", GF_LOG_DEBUG,
+                        gf_log (this->name, GF_LOG_DEBUG,
                                 "Error, Cannot Validate option :%s",
                                 *op_errstr);
                 } else {
-                        gf_log ("glsuterd", GF_LOG_DEBUG,
+                        gf_log (this->name, GF_LOG_DEBUG,
                                 "Error, Cannot Validate option");
                 }
         }
diff --git a/xlators/mgmt/glusterd/src/glusterd-volgen.c b/xlators/mgmt/glusterd/src/glusterd-volgen.c
index f15f57edc22..96df8585f65 100644
--- a/xlators/mgmt/glusterd/src/glusterd-volgen.c
+++ b/xlators/mgmt/glusterd/src/glusterd-volgen.c
@@ -109,30 +109,30 @@ struct volopt_map_entry {
 
 static struct volopt_map_entry glusterd_volopt_map[] = {
 
-        {"cluster.lookup-unhashed",              "cluster/distribute", NULL, NULL, NO_DOC, 0    },
-        {"cluster.min-free-disk",                "cluster/distribute", NULL, NULL, NO_DOC, 0    },
-	{"cluster.min-free-inodes",              "cluster/distribute", NULL, NULL, NO_DOC, 0    },
-
-        {"cluster.entry-change-log",             "cluster/replicate",  NULL, NULL, NO_DOC, 0     },
-        {"cluster.read-subvolume",               "cluster/replicate",  NULL, NULL, NO_DOC, 0    },
-        {"cluster.background-self-heal-count",   "cluster/replicate",  NULL, NULL, NO_DOC, 0    },
-        {"cluster.metadata-self-heal",           "cluster/replicate",  NULL, NULL, NO_DOC, 0     },
-        {"cluster.data-self-heal",               "cluster/replicate",  NULL, NULL, NO_DOC, 0     },
-        {"cluster.entry-self-heal",              "cluster/replicate",  NULL, NULL, NO_DOC, 0     },
-        {"cluster.self-heal-daemon",             "cluster/replicate",  "!self-heal-daemon" , NULL, NO_DOC, 0     },
-        {"cluster.strict-readdir",               "cluster/replicate",  NULL, NULL, NO_DOC, 0     },
-        {"cluster.self-heal-window-size",        "cluster/replicate",         "data-self-heal-window-size", NULL, DOC, 0},
-        {"cluster.data-change-log",              "cluster/replicate",  NULL, NULL, NO_DOC, 0     },
-        {"cluster.metadata-change-log",          "cluster/replicate",  NULL, NULL, NO_DOC, 0     },
-        {"cluster.data-self-heal-algorithm",     "cluster/replicate",         "data-self-heal-algorithm", NULL,DOC, 0},
+        {"cluster.lookup-unhashed",              "cluster/distribute", NULL, NULL, NO_DOC, 0},
+        {"cluster.min-free-disk",                "cluster/distribute", NULL, NULL, NO_DOC, 0},
+	{"cluster.min-free-inodes",              "cluster/distribute", NULL, NULL, NO_DOC, 0},
+
+        {"cluster.entry-change-log",             "cluster/replicate",  NULL, NULL, NO_DOC, 0},
+        {"cluster.read-subvolume",               "cluster/replicate",  NULL, NULL, NO_DOC, 0},
+        {"cluster.background-self-heal-count",   "cluster/replicate",  NULL, NULL, NO_DOC, 0},
+        {"cluster.metadata-self-heal",           "cluster/replicate",  NULL, NULL, NO_DOC, 0},
+        {"cluster.data-self-heal",               "cluster/replicate",  NULL, NULL, NO_DOC, 0},
+        {"cluster.entry-self-heal",              "cluster/replicate",  NULL, NULL, NO_DOC, 0},
+        {"cluster.self-heal-daemon",             "cluster/replicate",  "!self-heal-daemon", NULL, NO_DOC, 0},
+        {"cluster.strict-readdir",               "cluster/replicate",  NULL, NULL, NO_DOC, 0},
+        {"cluster.self-heal-window-size",        "cluster/replicate",  "data-self-heal-window-size", NULL, DOC, 0},
+        {"cluster.data-change-log",              "cluster/replicate",  NULL, NULL, NO_DOC, 0},
+        {"cluster.metadata-change-log",          "cluster/replicate",  NULL, NULL, NO_DOC, 0},
+        {"cluster.data-self-heal-algorithm",     "cluster/replicate",  "data-self-heal-algorithm", NULL,DOC, 0},
         {"cluster.quorum-type",                  "cluster/replicate",  "quorum-type", NULL, NO_DOC, 0},
         {"cluster.quorum-count",                 "cluster/replicate",  "quorum-count", NULL, NO_DOC, 0},
 
-        {"cluster.stripe-block-size",            "cluster/stripe",            "block-size", NULL, DOC, 0},
+        {"cluster.stripe-block-size",            "cluster/stripe",     "block-size", NULL, DOC, 0},
 
-        {VKEY_DIAG_LAT_MEASUREMENT,              "debug/io-stats",     "latency-measurement", "off", NO_DOC, 0      },
-        {"diagnostics.dump-fd-stats",            "debug/io-stats",     NULL, NULL, NO_DOC, 0     },
-        {VKEY_DIAG_CNT_FOP_HITS,                 "debug/io-stats",     "count-fop-hits", "off", NO_DOC, 0     },
+        {VKEY_DIAG_LAT_MEASUREMENT,              "debug/io-stats",     "latency-measurement", "off", NO_DOC, 0},
+        {"diagnostics.dump-fd-stats",            "debug/io-stats",     NULL, NULL, NO_DOC, 0},
+        {VKEY_DIAG_CNT_FOP_HITS,                 "debug/io-stats",     "count-fop-hits", "off", NO_DOC, 0},
 
         {"diagnostics.brick-log-level",          "debug/io-stats",     "!brick-log-level", NULL, DOC, 0},
         {"diagnostics.client-log-level",         "debug/io-stats",     "!client-log-level", NULL, DOC, 0},
@@ -143,28 +143,26 @@ static struct volopt_map_entry glusterd_volopt_map[] = {
         {"performance.cache-min-file-size",      "performance/io-cache",      "min-file-size", NULL, DOC, 0},
         {"performance.cache-refresh-timeout",    "performance/io-cache",      "cache-timeout", NULL, DOC, 0},
         {"performance.cache-priority",           "performance/io-cache",      "priority", NULL, DOC, 0},
-        {"performance.cache-size",               "performance/io-cache",   NULL, NULL, NO_DOC, 0 },
-        {"performance.cache-size",               "performance/quick-read", NULL, NULL, NO_DOC, 0 },
-        {"performance.flush-behind",             "performance/write-behind",      "flush-behind", NULL, DOC, 0},
+        {"performance.cache-size",               "performance/io-cache",      NULL, NULL, NO_DOC, 0 },
+        {"performance.cache-size",               "performance/quick-read",    NULL, NULL, NO_DOC, 0 },
+        {"performance.flush-behind",             "performance/write-behind",  "flush-behind", NULL, DOC, 0},
 
         {"performance.io-thread-count",          "performance/io-threads",    "thread-count", DOC, 0},
-
-        {"performance.disk-usage-limit",         "performance/quota",   NULL, NULL, NO_DOC, 0    },
-        {"performance.min-free-disk-limit",      "performance/quota",   NULL, NULL, NO_DOC, 0    },
-
+        {"performance.disk-usage-limit",         "performance/quota",         NULL, NULL, NO_DOC, 0},
+        {"performance.min-free-disk-limit",      "performance/quota",         NULL, NULL, NO_DOC, 0},
         {"performance.write-behind-window-size", "performance/write-behind",  "cache-size", NULL, DOC},
+        {"performance.read-ahead-page-count",    "performance/read-ahead",    "page-count", NULL, DOC},
 
-        {"performance.read-ahead-page-count",    "performance/read-ahead",  "page-count", NULL, DOC},
-
-        {"network.frame-timeout",                "protocol/client",    NULL, NULL, NO_DOC, 0     },
-        {"network.ping-timeout",                 "protocol/client",    NULL, NULL, NO_DOC, 0     },
-        {"network.inode-lru-limit",              "protocol/server",    NULL, NULL, NO_DOC, 0     },
+        {"network.frame-timeout",                "protocol/client",           NULL, NULL, NO_DOC, 0},
+        {"network.ping-timeout",                 "protocol/client",           NULL, NULL, NO_DOC, 0},
+        {"network.tcp-window-size",              "protocol/client",           NULL, NULL, NO_DOC, 0},
 
+        {"network.tcp-window-size",              "protocol/server",           NULL, NULL, NO_DOC, 0},
+        {"network.inode-lru-limit",              "protocol/server",           NULL, NULL, NO_DOC, 0},
         {"auth.allow",                           "protocol/server",           "!server-auth", "*", DOC, 0},
         {"auth.reject",                          "protocol/server",           "!server-auth", NULL, DOC, 0},
-
-        {"transport.keepalive",                   "protocol/server",           "transport.socket.keepalive", NULL, NO_DOC, 0},
-        {"server.allow-insecure",                 "protocol/server",          "rpc-auth-allow-insecure", NULL, NO_DOC, 0},
+        {"transport.keepalive",                  "protocol/server",           "transport.socket.keepalive", NULL, NO_DOC, 0},
+        {"server.allow-insecure",                "protocol/server",           "rpc-auth-allow-insecure", NULL, NO_DOC, 0},
 
         {"performance.write-behind",             "performance/write-behind",  "!perf", "on", NO_DOC, 0},
         {"performance.read-ahead",               "performance/read-ahead",    "!perf", "on", NO_DOC, 0},
@@ -2108,7 +2106,7 @@ volgen_graph_build_clients (volgen_graph_t *graph, glusterd_volinfo_t *volinfo,
                                        &client_type);
 
                 if (!ret && client_type == GF_CLIENT_TRUSTED) {
-
+                        str = NULL;
                         str = glusterd_auth_get_username (volinfo);
                         if (str) {
                                 ret = xlator_set_option (xl, "username",
diff --git a/xlators/mgmt/glusterd/src/glusterd.c b/xlators/mgmt/glusterd/src/glusterd.c
index f5ec6597663..d7912c95fa9 100644
--- a/xlators/mgmt/glusterd/src/glusterd.c
+++ b/xlators/mgmt/glusterd/src/glusterd.c
@@ -952,12 +952,12 @@ init (xlator_t *this)
         conf->valgrind = _gf_false;
         ret = dict_get_str (this->options, "brick-with-valgrind", &valgrind_str);
         if (ret < 0) {
-                gf_log (THIS->name, GF_LOG_ERROR,
+                gf_log (this->name, GF_LOG_DEBUG,
                         "cannot get brick-with-valgrind value");
         }
         if (valgrind_str) {
                 if (gf_string2boolean (valgrind_str, &(conf->valgrind))) {
-                        gf_log (THIS->name, GF_LOG_WARNING,
+                        gf_log (this->name, GF_LOG_WARNING,
                                 "brick-with-valgrind value not a boolean string");
                 }
         }
@@ -1126,8 +1126,10 @@ struct volume_options options[] = {
         { .key = {GEOREP"-log-group"},
           .type = GF_OPTION_TYPE_ANY,
         },
+#ifdef DEBUG
         { .key = {"brick-with-valgrind"},
           .type = GF_OPTION_TYPE_BOOL,
         },
+#endif
         { .key   = {NULL} },
 };
diff --git a/xlators/protocol/client/src/client.c b/xlators/protocol/client/src/client.c
index d93a99cff6e..8fadc040010 100644
--- a/xlators/protocol/client/src/client.c
+++ b/xlators/protocol/client/src/client.c
@@ -2586,5 +2586,10 @@ struct volume_options options[] = {
         { .key   = {"grace-timeout"},
           .type  = GF_OPTION_TYPE_INT
         },
+        {.key  = {"tcp-window-size"},
+         .type = GF_OPTION_TYPE_SIZET,
+         .min  = GF_MIN_SOCKET_WINDOW_SIZE,
+         .max  = GF_MAX_SOCKET_WINDOW_SIZE
+        },
         { .key   = {NULL} },
 };
diff --git a/xlators/protocol/client/src/client.h b/xlators/protocol/client/src/client.h
index f7267b44a67..f5f3aa1fec8 100644
--- a/xlators/protocol/client/src/client.h
+++ b/xlators/protocol/client/src/client.h
@@ -35,6 +35,8 @@
 #define CLIENT_CMD_CONNECT    "trusted.glusterfs.client-connect"
 #define CLIENT_CMD_DISCONNECT "trusted.glusterfs.client-disconnect"
 #define CLIENT_DUMP_LOCKS     "trusted.glusterfs.clientlk-dump"
+#define GF_MAX_SOCKET_WINDOW_SIZE  (1 * GF_UNIT_MB)
+#define GF_MIN_SOCKET_WINDOW_SIZE  (0)
 
 #define CLIENT_GET_REMOTE_FD(conf, fd, remote_fd, op_errno, label)      \
         do {                                                            \
diff --git a/xlators/protocol/server/src/server.c b/xlators/protocol/server/src/server.c
index 0f9195f9e66..bfac42a2799 100644
--- a/xlators/protocol/server/src/server.c
+++ b/xlators/protocol/server/src/server.c
@@ -1100,5 +1100,10 @@ struct volume_options options[] = {
         {.key  = {"grace-timeout"},
          .type = GF_OPTION_TYPE_INT,
         },
+        {.key  = {"tcp-window-size"},
+         .type = GF_OPTION_TYPE_SIZET,
+         .min  = GF_MIN_SOCKET_WINDOW_SIZE,
+         .max  = GF_MAX_SOCKET_WINDOW_SIZE
+        },
         { .key   = {NULL} },
 };
diff --git a/xlators/protocol/server/src/server.h b/xlators/protocol/server/src/server.h
index 091a02ccba2..4c38c19719d 100644
--- a/xlators/protocol/server/src/server.h
+++ b/xlators/protocol/server/src/server.h
@@ -32,6 +32,8 @@
 
 #define DEFAULT_BLOCK_SIZE         4194304   /* 4MB */
 #define DEFAULT_VOLUME_FILE_PATH   CONFDIR "/glusterfs.vol"
+#define GF_MAX_SOCKET_WINDOW_SIZE  (1 * GF_UNIT_MB)
+#define GF_MIN_SOCKET_WINDOW_SIZE  (0)
 
 typedef struct _server_state server_state_t;