Add flag-based setting for the distribution buffer busy limit

Id: OTP-8912

This patch creates a new family of flags with the "+z" prefix.  It
further creates a new configuration option called "dbbl" (which is the
first letter of the name dist_buf_busy_limit).  Example usage of this
flag would be "+zdbbl 1048576".

This patch creates an adjustable buffer limit for the amount of data
that may be buffered by the erlang distribution code (in dist.c
specifically). Before this patch, this hard-coded constant was used:

    #define ERTS_DE_BUSY_LIMIT (128*1024)

When large binaries are transmitted between nodes (or simply a lot of
medium-sized binaries), it is very easy to hit the old 128KB limit.
Processes that use the erlang:system_monitor() BIF to monitor system
events can be spammed by {monitor, busy_dist_port, ...} message tuples
at rates of tens to even hundreds of messages/second.

A larger buffer limit will allow processes to buffer more outgoing
messages over the distribution. When the buffer limit has been
reached, sending processes will be suspended until the buffer size has
shrunk. The buffer limit is per distribution channel. A higher limit
will give lower latency and higher throughput at the expense of
higher memory usage.

A variation of this patch has been in commercial production use in at
least two companies that the author is aware of.  Larger buffer values
can reduce the number of {monitor, busy_dist_port, ...} system
messages drastically, lower overall messaging latencies, and prevent
false timeouts and 'nodedown' messages in extremely busy Mnesia systems.

Test suite: there are two tests:

   a. In erlexec_SUITE.erl to test basic set & get of the value
   b. In distribution_SUITE.erl, to verify that setting +zdbbl very
      low will actually change behavior.

4 files changed, 36 insertions, 7 deletions

diff --git a/erts/emulator/beam/dist.c b/erts/emulator/beam/dist.c
index 16b6aeac3f..4497e17d79 100644
--- a/erts/emulator/beam/dist.c
+++ b/erts/emulator/beam/dist.c
@@ -97,6 +97,8 @@ dist_msg_dbg(ErtsDistExternal *edep, char *what, byte *buf, int sz)
 #define PASS_THROUGH 'p'        /* This code should go */
 
 int erts_is_alive; /* System must be blocked on change */
+int erts_dist_buf_busy_limit;
+
 
 /* distribution trap functions */
 Export* dsend2_trap = NULL;
@@ -1453,8 +1455,6 @@ int erts_net_message(Port *prt,
     return -1;
 }
 
-#define ERTS_DE_BUSY_LIMIT (128*1024)
-
 static int
 dsig_send(ErtsDSigData *dsdp, Eterm ctl, Eterm msg, int force_busy)
 {
@@ -1540,7 +1540,7 @@ dsig_send(ErtsDSigData *dsdp, Eterm ctl, Eterm msg, int force_busy)
 	ErtsProcList *plp = NULL;
 	erts_smp_spin_lock(&dep->qlock);
 	dep->qsize += size_obuf(obuf);
-	if (dep->qsize >= ERTS_DE_BUSY_LIMIT)
+	if (dep->qsize >= erts_dist_buf_busy_limit)
 	    dep->qflgs |= ERTS_DE_QFLG_BUSY;
 	if (!force_busy && (dep->qflgs & ERTS_DE_QFLG_BUSY)) {
 	    erts_smp_spin_unlock(&dep->qlock);
@@ -1911,7 +1911,7 @@ erts_dist_command(Port *prt, int reds_limit)
 	ASSERT(dep->qsize >= obufsize);
 	dep->qsize -= obufsize;
 	obufsize = 0;
-	if (de_busy && !prt_busy && dep->qsize < ERTS_DE_BUSY_LIMIT) {
+	if (de_busy && !prt_busy && dep->qsize < erts_dist_buf_busy_limit) {
 	    ErtsProcList *suspendees;
 	    int resumed;
 	    suspendees = get_suspended_on_de(dep, ERTS_DE_QFLG_BUSY);
diff --git a/erts/emulator/beam/dist.h b/erts/emulator/beam/dist.h
index fa19c7fb45..28cdd05c3c 100644
--- a/erts/emulator/beam/dist.h
+++ b/erts/emulator/beam/dist.h
@@ -99,7 +99,8 @@ typedef struct {
 #define ERTS_DE_IS_CONNECTED(DEP) \
   (!ERTS_DE_IS_NOT_CONNECTED((DEP)))
 
-
+#define ERTS_DE_BUSY_LIMIT (128*1024)
+extern int erts_dist_buf_busy_limit;
 extern int erts_is_alive;
 
 /*
diff --git a/erts/emulator/beam/erl_bif_info.c b/erts/emulator/beam/erl_bif_info.c
index 40d8dc097c..801263ec26 100644
--- a/erts/emulator/beam/erl_bif_info.c
+++ b/erts/emulator/beam/erl_bif_info.c
@@ -2533,6 +2533,13 @@ BIF_RETTYPE system_info_1(BIF_ALIST_1)
 	BIF_RET(erts_nif_taints(BIF_P));
     } else if (ERTS_IS_ATOM_STR("reader_groups_map", BIF_ARG_1)) {
 	BIF_RET(erts_get_reader_groups_map(BIF_P));
+    } else if (ERTS_IS_ATOM_STR("dist_buf_busy_limit", BIF_ARG_1)) {
+	Uint hsz = 0;
+
+ 	(void) erts_bld_uint(NULL, &hsz, erts_dist_buf_busy_limit);
+	hp = hsz ? HAlloc(BIF_P, hsz) : NULL;
+	res = erts_bld_uint(&hp, NULL, erts_dist_buf_busy_limit);
+	BIF_RET(res);
     }
 
     BIF_ERROR(BIF_P, BADARG);
diff --git a/erts/emulator/beam/erl_init.c b/erts/emulator/beam/erl_init.c
index 4ae656a3ad..36fefc5cba 100644
--- a/erts/emulator/beam/erl_init.c
+++ b/erts/emulator/beam/erl_init.c
@@ -535,7 +535,8 @@ void erts_usage(void)
 
     erts_fprintf(stderr, "-W<i|w>     set error logger warnings mapping,\n");
     erts_fprintf(stderr, "            see error_logger documentation for details\n");
-
+    erts_fprintf(stderr, "-zdbbl size set the distribution buffer busy limit in kilobytes\n");
+    erts_fprintf(stderr, "            valid range is [1-%d]\n", INT_MAX/1024);
     erts_fprintf(stderr, "\n");
     erts_fprintf(stderr, "Note that if the emulator is started with erlexec (typically\n");
     erts_fprintf(stderr, "from the erl script), these flags should be specified with +.\n");
@@ -818,7 +819,7 @@ early_init(int *argc, char **argv) /*
     erl_sys_args(argc, argv);
 
     erts_ets_realloc_always_moves = 0;
-
+    erts_dist_buf_busy_limit = ERTS_DE_BUSY_LIMIT;
 }
 
 #ifndef ERTS_SMP
@@ -1346,6 +1347,26 @@ erl_start(int argc, char **argv)
 	    }
 	    break;
 
+	case 'z': {
+	    char *sub_param = argv[i]+2;
+	    int new_limit;
+
+	    if (has_prefix("dbbl", sub_param)) {
+		arg = get_arg(sub_param+4, argv[i+1], &i);
+		new_limit = atoi(arg);
+		if (new_limit < 1 || INT_MAX/1024 < new_limit) {
+		    erts_fprintf(stderr, "Invalid dbbl limit: %d\n", new_limit);
+		    erts_usage();
+		} else {
+		    erts_dist_buf_busy_limit = new_limit*1024;
+		}
+	    } else {
+		erts_fprintf(stderr, "bad -z option %s\n", argv[i]);
+		erts_usage();
+	    }
+	    break;
+        }
+
 	default:
 	    erts_fprintf(stderr, "%s unknown flag %s\n", argv[0], argv[i]);
 	    erts_usage();