Added a new configuration parameter to allow the administrator to con…

…trol

the erlang virtual machines garbage collection knobs when spawning a new
process to handle incoming connections. The option is called "process_options".
Its default value is the empty list which uses erlangs default options as
is the case now. This is useful for long lived connections which can generate
a lot of garbage if left untapped. Use with care :-)

include/yaws.hrl

@@ -77,7 +77,11 @@
                max_num_cached_files = 400,
                max_num_cached_bytes = 1000000,  %% 1 MEG
                max_size_cached_file = 8000,
-	       max_connections = nolimit, %% max number of TCP connections 
+               max_connections = nolimit, %% max number of TCP connections 
+               process_options = [],      %% Override default connection handler processes
+                                          %% spawn options for performance/memory tuning.
+                                          %% [] | [{fullsweep_after, Number}, {min_heap_size, Size}]
+                                          %% other options such as monitor, link are ignored.
                large_file_chunk_size = 10240,
                mnesia_dir = [],
                log_wrap_size = 10000000,  % wrap logs after 10M

man/yaws.conf.5

@@ -107,6 +107,15 @@ Set this value to control the maximum number of connections
 from HTTP clients into the server. This is implemented by closing
 the last socket if the limit threshold is reached.
 
+.TP
+\fB process_options = "[]" | "[{fullsweep_after, int()} | {min_heap_size, int()}]"\fR
+Override the garbage collection option parameters for processes
+which handle new connections. Useful for systems which expect long lived
+connections which handle a lot of data. The default value is erlangs
+default which does minimal garbage collection until the process dies.
+The value type is a quoted string which contains an erlang proplist.
+See erlangs erlang:spawn_opt/4 function for more details.
+
 .TP
 \fB log_wrap_size = Integer\fR
 The logs written by yaws are all wrap logs, the default value at the

scripts/yaws.conf.template

@@ -29,6 +29,13 @@ include_dir = %yawsdir%/examples/include
 # max number of connections from clients into the server
 max_connections = nolimit
 
+# Override the garbage collection option parameters for processes
+# which handle new connections. Useful for systems which expect long lived
+# connections which handle a lot of data. The default value is erlangs
+# default. Valid options are {fullsweep_after, X} and/or {min_heap_size, Y} where
+# X and Y are integers. See erlangs erlang:spawn_opt/4 function for more details.
+# The value type is a quoted string containing an erlang proplist
+process_options = "[]"
 
 # This is a debug variable, possible values are http | traffic | false
 # It is also possible to set the trace (possibly to the tty) while

src/yaws_config.erl

@@ -584,6 +584,15 @@ fload(FD, globals, GC, C, Cs, Lno, Chars) ->
                     {error, ?F("Expect integer at line ~w", [Lno])}
             end;
 
+        ["process_options", '=', POpts] ->
+            case parse_process_options(POpts) of
+                {ok, ProcList} ->
+                    fload(FD, globals, GC#gconf{process_options=ProcList},
+                          C, Cs, Lno+1, Next);
+                {error, Str} ->
+                    {error, ?F("~s at line ~w", [Str, Lno])}
+            end;
+
         ["log_wrap_size", '=', Int] ->
             case (catch list_to_integer(Int)) of
                 I when is_integer(I) ->
@@ -1467,6 +1476,39 @@ is_string_char([C|T]) ->
 is_special(C) ->
     lists:member(C, [$=, $<, $>, $,]).
 
+%% parse the argument string PLString which can either be the undefined atom
+%% or a proplist. Currently the only supported keys are fullsweep_after and
+%% min_heap_size. Any other key/values are ignored.
+parse_process_options(PLString) ->
+    case erl_scan:string(PLString ++ ".") of
+        {ok, PLTokens, _} ->
+            case erl_parse:parse_term(PLTokens) of
+                {ok, undefined} ->
+                    {ok, []};
+                {ok, []} ->
+                    {ok, []};
+                {ok, [Hd|_Tl]=PList} when is_atom(Hd); is_tuple(Hd) ->
+                    %% create new safe proplist of desired options
+                    {ok, proplists_int_copy([], PList, [fullsweep_after, min_heap_size])};
+                _ ->
+                    {error, "Expect undefined or proplist"}
+            end;
+        _ ->
+            {error, "Expect undefined or proplist"}
+    end.
+
+%% copy proplist integer values for the given keys from the 
+%% Src proplist to the Dest proplist. Ignored keys that are not
+%% found or have non-integer values. Returns the new Dest proplist.
+proplists_int_copy(Dest, _Src, []) ->
+    Dest;
+proplists_int_copy(Dest, Src, [Key|NextKeys]) ->
+    case proplists:get_value(Key, Src) of
+        Val when is_integer(Val) ->
+            proplists_int_copy([{Key, Val}|Dest], Src, NextKeys);
+        _ ->
+            proplists_int_copy(Dest, Src, NextKeys)
+    end.
 
 parse_soap_srv_mods(['<', Module, ',' , Handler, ',', WsdlFile, '>' | Tail], 
                     Ack) ->

src/yaws_server.erl

@@ -908,7 +908,14 @@ initial_acceptor(GS) ->
 
 
 acceptor(GS) ->
-    proc_lib:spawn_link(?MODULE, acceptor0, [GS, self()]).
+    case (GS#gs.gconf)#gconf.process_options of
+        [] ->
+            proc_lib:spawn_link(?MODULE, acceptor0, [GS, self()]);
+        Opts ->
+            %% as we tightly controlled what is set in options, we can blindly add "link" to
+            %% get a linked process as per default case and use the provided options.
+            proc_lib:spawn_opt(?MODULE, acceptor0, [GS, self()], [link | Opts])
+    end.
 acceptor0(GS, Top) ->
     ?TC([{record, GS, gs}]),
     put(gc, GS#gs.gconf),