Gracefully handle EMFILE

Implementing a tip from Marc Lehmann:
http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod#The_special_problem_of_accept_ing_wh

Keep an extra FD around for every server. When you hit EMFILE, destroy that
FD, accept a connection, close it; in this way you can clear the connection
queue and let people know that you're overload.

No more timeout needed.

1 files changed, 41 insertions, 20 deletions

diff --git a/lib/net.js b/lib/net.js
index f25012364..1e11c4aa2 100644
--- a/lib/net.js
+++ b/lib/net.js
@@ -922,36 +922,38 @@ function Server (/* [ options, ] listener */) {
 
   self.connections = 0;
 
-  self.paused = false;
-  self.pauseTimeout = 1000;
-
   self.allowHalfOpen = options.allowHalfOpen || false;
 
-  function pause () {
-    // We've hit the maximum file limit. What to do?
-    // Let's try again in 1 second.
-    self.watcher.stop();
-
-    // If we're already paused, don't do another timeout.
-    if (self.paused) return;
-
-    setTimeout(function () {
-      self.paused = false;
-      // Make sure we haven't closed in the interim
-      if (typeof self.fd  != 'number') return;
-      self.watcher.start();
-    }, self.pauseTimeout);
-  }
-
   self.watcher = new IOWatcher();
   self.watcher.host = self;
   self.watcher.callback = function () {
+    // Just in case we don't have a dummy fd.
+    if (!self._dummyFD) self._getDummyFD();
+
     while (self.fd) {
       try {
         var peerInfo = accept(self.fd);
       } catch (e) {
         if (e.errno == EMFILE) {
-          pause();
+          // Output a warning, but only at most every 5 seconds.
+          var now = new Date();
+          if (now - self._lastEMFILEWarning > 5000) {
+            console.error("(node) Hit max file limit. Increase 'ulimit -n'.");
+          }
+          self._lastEMFILEWarning = now;
+
+          // Gracefully reject pending clients by freeing up a file
+          // descriptor.
+          if (self._dummyFD) {
+            close(self._dummyFD);
+            self._dummyFD = null;
+            while (true) {
+              peerInfo = accept(self.fd);
+              if (!peerInfo) break;
+              close(peerInfo.fd);  
+            }
+            self._getDummyFD();
+          }
           return;
         }
         throw e;
@@ -1076,9 +1078,23 @@ Server.prototype._startWatcher = function () {
   this.emit("listening");
 };
 
+
+Server.prototype._getDummyFD = function () {
+  try {
+    this._dummyFD = socket("tcp");
+  } catch (e) {
+    this._dummyFD = null;
+  }
+};
+
+
 Server.prototype._doListen = function () {
   var self = this;
 
+  // Grab a dummy fd for EMFILE conditions.
+  self._getDummyFD();
+  self._lastEMFILEWarning = 0;
+
   try {
     bind(self.fd, arguments[0], arguments[1]);
   } catch (err) {
@@ -1120,6 +1136,11 @@ Server.prototype.close = function () {
   close(self.fd);
   self.fd = null;
 
+  if (this._dummyFD) {
+    close(this._dummyFD);
+    this._dummyFD = null;
+  }
+
   if (self.type === "unix") {
     fs.unlink(self.path, function () {
       self.emit("close");