--- loncom/Attic/lonc	2002/02/25 20:43:15	1.30
+++ loncom/Attic/lonc	2002/04/04 22:04:54	1.38
@@ -5,7 +5,7 @@
 # provides persistent TCP connections to the other servers in the network
 # through multiplexed domain sockets
 #
-# $Id: lonc,v 1.30 2002/02/25 20:43:15 www Exp $
+# $Id: lonc,v 1.38 2002/04/04 22:04:54 foxr Exp $
 #
 # Copyright Michigan State University Board of Trustees
 #
@@ -45,7 +45,7 @@
 # 12/20 Scott Harrison
 # YEAR=2002
 # 2/19/02,02/22/02,02/25/02 Gerd Kortemeyer
-# 
+# 3/07/02 Ron Fox 
 # based on nonforker from Perl Cookbook
 # - server who multiplexes without forking
 
@@ -57,12 +57,13 @@ use Socket;
 use Fcntl;
 use Tie::RefHash;
 use Crypt::IDEA;
-use Net::Ping;
+#use Net::Ping;
 use LWP::UserAgent();
 
 $status='';
 $lastlog='';
 $conserver='SHELL';
+$DEBUG = 0;			# Set to 1 for annoyingly complete logs.
 
 # -------------------------------- Set signal handlers to record abnormal exits
 
@@ -158,47 +159,53 @@ $SIG{HUP}=$SIG{USR1}='IGNORE';
 &status("Forking ...");
 
 foreach $thisserver (keys %hostip) {
-    if (&online($hostname{$thisserver})) {
+    #if (&online($hostname{$thisserver})) {
        make_new_child($thisserver);
-    }
+    #}
 }
 
 &logthis("Done starting initial servers");
 # ----------------------------------------------------- Install signal handlers
 
-$SIG{CHLD} = \&REAPER;
+
 $SIG{INT}  = $SIG{TERM} = \&HUNTSMAN;
 $SIG{HUP}  = \&HUPSMAN;
 $SIG{USR1} = \&USRMAN;
 
 # And maintain the population.
 while (1) {
-    &status("Sleeping");
-    sleep;                          # wait for a signal (i.e., child's death)
+    my $deadpid = wait;		# Wait for the next child to die.
                                     # See who died and start new one
     &status("Woke up");
     my $skipping='';
-    foreach $thisserver (keys %hostip) {
-        if (!$childpid{$thisserver}) {
-	    if (($childatt{$thisserver}<$childmaxattempts) &&
-                (&online($hostname{$thisserver}))) {
-	       $childatt{$thisserver}++;
-               &logthis(
-   "<font color=yellow>INFO: Trying to reconnect for $thisserver "
-  ."($childatt{$thisserver} of $childmaxattempts attempts)</font>"); 
-               make_new_child($thisserver);
-	   } else {
-               $skipping.=$thisserver.' ';
-           } 
-               
-        }       
-    }
-    if ($skipping) { 
-       &logthis("<font color=blue>WARNING: Skipped $skipping</font>");
+
+    if(exists($children{$deadpid})) {
+
+	$thisserver = $children{$deadpid}; # Look name of dead guy's peer.
+
+	delete($children{$deadpid}); # Get rid of dead hash entry.
+
+	if($childatt{$thisserver} < $childmaxattempts) {
+	    $childatt{$thisserver}++;
+	    &logthis(
+	       "<font color=yellow>INFO: Trying to reconnect for $thisserver "
+            ."($childatt{$thisserver} of $childmaxattempts attempts)</font>"); 
+	    make_new_child($thisserver);
+	
+	}
+	else {
+	    $skipping .= $thisserver.' ';
+	}
+	if($skipping) {
+	    &logthis("<font color=blue>WARNING: Skipped $skipping</font>");
+  
+	}
     }
+
 }
 
 
+
 sub make_new_child {
    
     $newserver=shift;
@@ -217,7 +224,7 @@ sub make_new_child {
         sigprocmask(SIG_UNBLOCK, $sigset)
             or die "Can't unblock SIGINT for fork: $!\n";
         $children{$pid} = $newserver;
-        $childpid{$conserver} = $pid;
+        $childpid{$newserver} = $pid;
         return;
     } else {
         $conserver=$newserver;
@@ -238,9 +245,10 @@ unlink($port);
 # -------------------------------------------------------------- Open other end
 
 &openremote($conserver);
-
+	&logthis("<font color=green> Connection to $conserver open </font>");
 # ----------------------------------------- We're online, send delayed messages
     &status("Checking for delayed messages");
+
     my @allbuffered;
     my $path="$perlvar{'lonSockDir'}/delayed";
     opendir(DIRHANDLE,$path);
@@ -250,7 +258,7 @@ unlink($port);
     foreach (@allbuffered) {
         &status("Sending delayed: $_");
         $dfname="$path/$_";
-        &logthis('Sending '.$dfname);
+        if($DEBUG) { &logthis('Sending '.$dfname); }
         my $wcmd;
         {
          my $dfh=IO::File->new($dfname);
@@ -271,17 +279,8 @@ unlink($port);
             }
             $cmd="enc:$cmdlength:$encrequest\n";
         }
-    $SIG{ALRM}=sub { die "timeout" };
-    $SIG{__DIE__}='DEFAULT';
-    eval {
-        alarm(60);
-        print $remotesock "$cmd\n";
-        $answer=<$remotesock>;
+	$answer = londtransaction($remotesock, $cmd, 60);
 	chomp($answer);
-        alarm(0);
-    };
-    $SIG{ALRM}='DEFAULT';
-    $SIG{__DIE__}=\&catchexception;
 
         if (($answer ne '') && ($@!~/timeout/)) {
 	    unlink("$dfname");
@@ -289,6 +288,7 @@ unlink($port);
             &logperm("S:$conserver:$bcmd");
         }        
     }
+	if($DEBUG) { &logthis("<font color=green> Delayed transactions sent"); }
 
 # ------------------------------------------------------- Listen to UNIX socket
 &status("Opening socket");
@@ -300,11 +300,11 @@ unless (
        my $st=120+int(rand(240));
        &logthis(
          "<font color=blue>WARNING: ".
-         "Can't make server socket ($st secs): $@</font>");
+         "Can't make server socket ($st secs):  .. exiting</font>");
        sleep($st);
        exit; 
      };
-
+   
 # -----------------------------------------------------------------------------
 
 &logthis("<font color=green>$conserver online</font>");
@@ -314,128 +314,308 @@ unless (
 %inbuffer  = ();
 %outbuffer = ();
 %ready     = ();
+%servers   = ();	# To be compatible with make filevector.  indexed by
+			# File ids, values are sockets.
+			# note that the accept socket is omitted.
 
 tie %ready, 'Tie::RefHash';
 
-nonblock($server);
-$select = IO::Select->new($server);
+# nonblock($server);
+# $select = IO::Select->new($server);
 
 # Main loop: check reads/accepts, check writes, check ready to process
+
+status("Main loop");
 while (1) {
     my $client;
     my $rv;
     my $data;
 
-    # check for new information on the connections we have
+    my $infdset;		# bit vec of fd's to select on input.
 
-    # anything to read or accept?
-    foreach $client ($select->can_read(0.1)) {
+    my $outfdset;		# Bit vec of fd's to select on output.
 
-        if ($client == $server) {
-            # accept a new connection
-            &status("Accept new connection: $conserver");
-            $client = $server->accept();
-            $select->add($client);
-            nonblock($client);
-        } else {
-            # read data
-            $data = '';
-            $rv   = $client->recv($data, POSIX::BUFSIZ, 0);
-
-            unless (defined($rv) && length $data) {
-                # This would be the end of file, so close the client
-                delete $inbuffer{$client};
-                delete $outbuffer{$client};
-                delete $ready{$client};
-
-                &status("Idle");
-                $select->remove($client);
-                close $client;
-                next;
-            }
 
-            $inbuffer{$client} .= $data;
+    $infdset = MakeFileVector(\%servers);
+    $outfdset= MakeFileVector(\%outbuffer);
+    vec($infdset, $server->fileno, 1) = 1;
+    if($DEBUG) {
+	&logthis("Adding ".$server->fileno.
+		 " to input select vector (listner)".
+		 unpack("b*",$infdset)."\n");
+    }
+    DoSelect(\$infdset, \$outfdset); # Wait for input.
+    if($DEBUG) {
+	&logthis("Doselect completed!");
+	&logthis("ins = ".unpack("b*",$infdset)."\n");
+	&logthis("outs= ".unpack("b*",$outfdset)."\n");
+		 
+    }
 
-            # test whether the data in the buffer or the data we
-            # just read means there is a complete request waiting
-            # to be fulfilled.  If there is, set $ready{$client}
-            # to the requests waiting to be fulfilled.
-            while ($inbuffer{$client} =~ s/(.*\n)//) {
-                push( @{$ready{$client}}, $1 );
-            }
-        }
+    # Checkfor new connections:
+    if (vec($infdset, $server->fileno, 1)) {
+	if($DEBUG) {
+	    &logthis("New connection established");
+	}
+	# accept a new connection
+	&status("Accept new connection: $conserver");
+	$client = $server->accept();
+	if($DEBUG) {
+	    &logthis("New client fd = ".$client->fileno."\n");
+	}
+	$servers{$client->fileno} = $client;
+	nonblock($client);
     }
+    HandleInput($infdset, \%servers, \%inbuffer, \%outbuffer, \%ready);
+    HandleOutput($outfdset, \%servers, \%outbuffer, \%inbuffer,
+		 \%ready);
+# -------------------------------------------------------- Wow, connection lost
 
-    # Any complete requests to process?
-    foreach $client (keys %ready) {
-        handle($client);
+}
+   
     }
+}
 
-    # Buffers to flush?
-    foreach $client ($select->can_write(1)) {
-        # Skip this client if we have nothing to say
-        next unless exists $outbuffer{$client};
+# ------------------------------------------------------- End of make_new_child
 
-        $rv = $client->send($outbuffer{$client}, 0);
 
-      unless ($outbuffer{$client} eq "con_lost\n") {
-        unless (defined $rv) {
-            # Whine, but move on.
-            &logthis("I was told I could write, but I can't.\n");
-            next;
-        }
-        $errno=$!;
-        if (($rv == length $outbuffer{$client}) ||
-            ($errno == POSIX::EWOULDBLOCK) || ($errno == 0)) {
-            substr($outbuffer{$client}, 0, $rv) = '';
-            delete $outbuffer{$client} unless length $outbuffer{$client};
-        } else {
-            # Couldn't write all the data, and it wasn't because
-            # it would have blocked.  Shutdown and move on.
+#
+#  Make a vector of file descriptors to wait for in a select.
+#  parameters:
+#     \%fdhash  -reference to a hash which has IO::Socket's as indices.  
+#                We only care about the indices, not the values.
+#  A select vector is created from all indices of the hash.
+
+sub MakeFileVector
+{
+    my $fdhash = shift;
+    my $selvar = "";
+
+    foreach $socket (keys %$fdhash) {
+	if($DEBUG) {
+	    &logthis("Adding  ".$socket.
+		     "to select vector. (client)\n");
+	}
+	vec($selvar, $socket, 1) = 1;
+    }
+    return $selvar;
+}
 
-	    &logthis("Dropping data with ".$errno.": ".
-                     length($outbuffer{$client}).", $rv");
 
-            delete $inbuffer{$client};
-            delete $outbuffer{$client};
-            delete $ready{$client};
-
-            $select->remove($client);
-            close($client);
-            next;
-        }
-      } else {
-# -------------------------------------------------------- Wow, connection lost
-         &logthis(
-     "<font color=red>CRITICAL: Closing connection</font>");
-	 &status("Connection lost");
-         $remotesock->shutdown(2);
-         &logthis("Attempting to open new connection");
-         &openremote($conserver);          
-      }
+#
+#  HandleOutput:
+#    Processes output on a buffered set of file descriptors which are
+#    ready to be read.
+#  Parameters:
+#    $selvector - Vector of file descriptors which are writable.
+#    \%sockets  - Vector of socket references indexed by socket.
+#    \%buffers  - Reference to a hash containing output buffers.
+#                 Hashes are indexed by sockets.  The file descriptors of some
+#                 of those sockets will be present in $selvector.
+#                 For each one of those, we will attempt to write the output
+#                 buffer to the socket.  Note that we will assume that
+#                 the sockets are being run in non blocking mode.
+#   \%inbufs    - Reference to hash containing input buffers.
+#   \%readys    - Reference to hash containing flags for items with complete
+#                 requests.
+#
+sub HandleOutput
+{
+    my $selvector = shift;
+    my $sockets   = shift;
+    my $buffers   = shift;
+    my $inbufs    = shift;
+    my $readys    = shift;
+    my $sock;
+
+    if($DEBUG) {
+	&logthis("HandleOutput entered\n");
     }
-   
+
+    foreach $sock (keys %$sockets) {
+	my $socket = $sockets->{$sock};
+	if(vec($selvector, $sock, 1)) { # $socket is writable.
+	    if($DEBUG) {
+		&logthis("Sending $buffers->{$sock} \n");
+	    }
+	    my $rv = $socket->send($buffers->{$sock}, 0);
+	    $errno = $!;
+	    unless ($buffers->{$sock} eq "con_lost\n") {
+		unless (defined $rv) { # Write failed... could be EINTR
+		    unless ($errno == POSIX::EINTR) {
+			&logthis("Write failed on writable socket");
+		    }		# EINTR is not an error .. just retry.
+		    next;
+		}
+		if( ($rv == length $buffers->{$sock})    ||
+		    ($errno == POSIX::EWOULDBLOCK)       ||
+		    ($errno == POSIX::EAGAIN)            || # same as above.
+		    ($errno == POSIX::EINTR)             || # signal during IO
+		    ($errno == 0)) {
+		    substr($buffers->{$sock}, 0, $rv)=""; # delete written part
+		    delete $buffers->{$sock} unless length $buffers->{$sock};
+		} else {
+		    # For some reason the write failed with an error code
+		    # we didn't look for.  Shutdown the socket.
+		    &logthis("Unable to write data with ".$errno.": ".
+			     "Dropping data: ".length($buffers->{$sock}).
+			     ", $rv");
+		    #
+		    # kill off the buffers in the hash:
+
+		    delete $buffers->{$sock};
+		    delete $inbufs->{$sock};
+		    delete $readys->{$sock};
+
+		    close($socket); # Close the client socket.
+		    next;
+		}
+	    } else {		# Kludgy way to mark lond connection lost.
+		&logthis(
+		 "<font color=red>CRITICAL lond connection lost</font>");
+		status("Connection lost");
+		$remotesock->shutdown(2);
+		&logthis("Attempting to open a new connection");
+		&openremote($conserver);
+	    }
+		   
+	}
+    }
+
 }
+#
+#   HandleInput - Deals with input on client sockets.
+#                 Each socket has an associated input buffer.
+#                 For each readable socket, the currently available
+#                 data is appended to this buffer.
+#                 If necessary, the buffer is created.
+#                 On various failures, we may shutdown the client.
+#  Parameters:
+#     $selvec   - Vector of readable sockets.
+#     \%sockets - Refers to the  Hash of sockets indexed by sockets.  
+#                 Each of these may or may not have it's fd bit set 
+#                 in the $selvec.
+#     \%ibufs   - Refers to the hash of input buffers indexed by socket.
+#     \%obufs   - Hash of output buffers indexed by socket. 
+#     \%ready   - Hash of ready flags indicating the existence of a completed
+#                 Request.
+sub HandleInput 
+{
+
+    # Marshall the parameters.   Note that the hashes are actually
+    # references not values.
+
+    my $selvec  = shift;
+    my $sockets = shift;
+    my $ibufs   = shift;
+    my $obufs   = shift;
+    my $ready   = shift;
+    my $sock;
+
+    if($DEBUG) {
+	&logthis("Entered HandleInput\n");
+    }
+    foreach $sock (keys %$sockets) {
+	my $socket = $sockets->{$sock};
+	if(vec($selvec, $sock, 1)) { # Socket which is readable.
+
+	    #  Attempt to read the data and do error management.
+	    my $data = '';
+	    my $rv = $socket->recv($data, POSIX::BUFSIZ, 0);
+	    if($DEBUG) {
+		&logthis("Received $data from socket");
+	    }
+	    unless (defined($rv) && length $data) {
+
+		# Read an end of file.. this is a disconnect from the peer.
+
+		delete $sockets->{$sock};
+		delete $ibufs->{$sock};
+		delete $obufs->{$sock};
+		delete $ready->{$sock};
+
+		status("Idle");
+		close $socket;
+		next;
+	    }
+	    #  Append the read data to the input buffer. If the buffer
+	    # now contains a \n the request is complete and we can 
+	    # mark this in the $ready hash (one request for each \n.)
+
+	    $ibufs->{$sock} .= $data;
+	    while($ibufs->{$sock} =~ s/(.*\n)//) {
+		push(@{$ready->{$sock}}, $1);
+	    }
+	    
+	}
+    }
+    #  Now handle any requests which are ready:
+
+    foreach $client (keys %ready) {
+	handle($client);
+    }
 }
 
-# ------------------------------------------------------- End of make_new_child
+# DoSelect:  does a select with no timeout.  On signal (errno == EINTR), 
+#            the select is retried until there are items in the returned
+#            vectors.  
+#
+# Parameters:
+#   \$readvec   - Reference to a vector of file descriptors to 
+#                 check for readability.
+#   \$writevec  - Reference to a vector of file descriptors to check for
+#                 writability.
+#  On exit, the referents are modified with vectors indicating which 
+#  file handles are readable/writable.
+#
+sub DoSelect {
+    my $readvec = shift;
+    my $writevec= shift;
+    my $outs;
+    my $ins;
+
+    while (1) {
+	my $nfds = select( $ins = $$readvec, $outs = $$writevec, undef, undef);
+	if($nfds) {
+	    if($DEBUG) {
+		&logthis("select exited with ".$nfds." fds\n");
+		&logthis("ins = ".unpack("b*",$ins).
+			 " readvec = ".unpack("b*",$$readvec)."\n");
+		&logthis("outs = ".unpack("b*",$outs).
+			 " writevec = ".unpack("b*",$$writevec)."\n");
+	    }
+	    $$readvec  = $ins;
+	    $$writevec = $outs;
+	    return;
+	} else {
+	    if($DEBUG) {
+		&logthis("Select exited with no bits set in mask\n");
+	    }
+	    die "Select failed" unless $! == EINTR;
+	}
+    }
+}
 
 # handle($socket) deals with all pending requests for $client
+#
 sub handle {
     # requests are in $ready{$client}
     # send output to $outbuffer{$client}
     my $client = shift;
     my $request;
-
     foreach $request (@{$ready{$client}}) {
 # ============================================================= Process request
         # $request is the text of the request
         # put text of reply into $outbuffer{$client}
 # ------------------------------------------------------------ Is this the end?
+	chomp($request);
+	if($DEBUG) {
+     &logthis("<font color=green> Request $request processing starts</font>");
+        }
         if ($request eq "close_connection_exit\n") {
 	    &status("Request close connection");
            &logthis(
-     "<font color=red>CRITICAL: Request Close Connection</font>");
+     "<font color=red>CRITICAL: Request Close Connection ... exiting</font>");
            $remotesock->shutdown(2);
            $server->close();
            exit;
@@ -452,27 +632,19 @@ sub handle {
                 $encrequest.=
                     unpack("H16",$cipher->encrypt(substr($cmd,$encidx,8)));
             }
-            $request="enc:$cmdlength:$encrequest\n";
+            $request="enc:$cmdlength:$encrequest";
         }
 # --------------------------------------------------------------- Main exchange
-    $SIG{ALRM}=sub { die "timeout" };
-    $SIG{__DIE__}='DEFAULT';
-    eval {
-        alarm(300);
-        &status("Sending: $request");
-        print $remotesock "$request";
-        &status("Waiting for reply from $conserver: $request");
-        $answer=<$remotesock>;
-        &status("Received reply: $request");
-        alarm(0);
-    };
-    if ($@=~/timeout/) { 
-       $answer='';
-       &logthis(
-        "<font color=red>CRITICAL: Timeout: $request</font>");
-    }  
-    $SIG{ALRM}='DEFAULT';
-    $SIG{__DIE__}=\&catchexception;
+	$answer = londtransaction($remotesock, $request, 300);
+
+	if($DEBUG) { 
+	    &logthis("<font color=green> Request data exchange complete");
+	}
+	if ($@=~/timeout/) { 
+	    $answer='';
+	    &logthis(
+		     "<font color=red>CRITICAL: Timeout: $request</font>");
+	}  
 
 
         if ($answer) {
@@ -488,20 +660,27 @@ sub handle {
 	      $answer=substr($answer,0,$cmdlength);
 	      $answer.="\n";
 	   }
+	   if($DEBUG) {
+	       &logthis("sending $answer to client\n");
+	   }
            $outbuffer{$client} .= $answer;
         } else {
            $outbuffer{$client} .= "con_lost\n";
         }
 
      &status("Completed: $request");
-
+	if($DEBUG) {
+	    &logthis("<font color=green> Request processing complete</font>");
+	}
 # ===================================================== Done processing request
     }
     delete $ready{$client};
 # -------------------------------------------------------------- End non-forker
+    if($DEBUG) {
+	&logthis("<font color=green> requests for child handled</font>");
+    }
 }
 # ---------------------------------------------------------- End make_new_child
-}
 
 # nonblock($socket) puts socket into nonblocking mode
 sub nonblock {
@@ -522,6 +701,7 @@ sub openremote {
     my $conserver=shift;
 
 &status("Opening TCP");
+    my $st=120+int(rand(240)); # Sleep before opening:
 
 unless (
   $remotesock = IO::Socket::INET->new(PeerAddr => $hostip{$conserver},
@@ -529,32 +709,24 @@ unless (
                                       Proto    => "tcp",
                                       Type     => SOCK_STREAM)
    ) { 
-       my $st=120+int(rand(240));
+
        &logthis(
-"<font color=blue>WARNING: Couldn't connect ($st secs): $@</font>");
+"<font color=blue>WARNING: Couldn't connect to $conserver ($st secs): </font>");
        sleep($st);
        exit; 
      };
 # ----------------------------------------------------------------- Init dialog
 
+&logthis("<font color=green>INFO Connected to $conserver, initing </font>");
 &status("Init dialogue: $conserver");
 
-     $SIG{ALRM}=sub { die "timeout" };
-     $SIG{__DIE__}='DEFAULT';
-     eval {
-         alarm(60);
-print $remotesock "init\n";
-$answer=<$remotesock>;
-print $remotesock "$answer";
-$answer=<$remotesock>;
-chomp($answer);
-          alarm(0);
-     };
-     $SIG{ALRM}='DEFAULT';
-     $SIG{__DIE__}=\&catchexception;
+    $answer = londtransaction($remotesock, "init", 60);
+    chomp($answer);
+    $answer = londtransaction($remotesock, $answer, 60);
+    chomp($answer);
  
      if ($@=~/timeout/) {
-	 &logthis("Timed out during init");
+	 &logthis("Timed out during init.. exiting");
          exit;
      }
 
@@ -597,7 +769,7 @@ if ($cipher=new IDEA $cipherkey) {
    sleep($st);
    exit;
 }
-
+    &logthis("<font color=green> Remote open success </font>");
 }
 
 
@@ -610,21 +782,21 @@ sub catchexception {
     chomp($signal);
     &logthis("<font color=red>CRITICAL: "
      ."ABNORMAL EXIT. Child $$ for server [$wasserver] died through "
-     ."\"$signal\" with parameter [$@]</font>");
-    die($@);
+     ."\"$signal\" with parameter </font>");
+    die("Signal abend");
 }
 
 # -------------------------------------- Routines to see if other box available
 
-sub online {
-    my $host=shift;
-    &status("Pinging ".$host);
-    my $p=Net::Ping->new("tcp",20);
-    my $online=$p->ping("$host");
-    $p->close();
-    undef ($p);
-    return $online;
-}
+#sub online {
+#    my $host=shift;
+#    &status("Pinging ".$host);
+#    my $p=Net::Ping->new("tcp",20);
+#    my $online=$p->ping("$host");
+#    $p->close();
+#    undef ($p);
+#    return $online;
+#}
 
 sub connected {
     my ($local,$remote)=@_;
@@ -635,7 +807,7 @@ sub connected {
     unless ($hostname{$local}) { return 'local_unknown'; }
     unless ($hostname{$remote}) { return 'remote_unknown'; }
 
-    unless (&online($hostname{$local})) { return 'local_offline'; }
+    #unless (&online($hostname{$local})) { return 'local_offline'; }
 
     my $ua=new LWP::UserAgent;
     
@@ -654,17 +826,6 @@ sub connected {
 }
 
 
-sub REAPER {                        # takes care of dead children
-    $SIG{CHLD} = \&REAPER;
-    my $pid = wait;
-    my $wasserver=$children{$pid};
-    &logthis("<font color=red>CRITICAL: "
-     ."Child $pid for server $wasserver died ($childatt{$wasserver})</font>");
-    delete $children{$pid};
-    delete $childpid{$wasserver};
-    my $port = "$perlvar{'lonSockDir'}/$wasserver";
-    unlink($port);
-}
 
 sub hangup {
     foreach (keys %children) {
@@ -725,15 +886,8 @@ sub subreply {
        or return "con_lost";
 
 
-    $SIG{ALRM}=sub { die "timeout" };
-    $SIG{__DIE__}='DEFAULT';
-    eval {
-     alarm(10);
-     print $sclient "$cmd\n";
-     $answer=<$sclient>;
-     chomp($answer);
-     alarm(0);
-    };
+    $answer = londtransaction($sclient, $cmd, 10);
+
     if ((!$answer) || ($@=~/timeout/)) { $answer="con_lost"; }
     $SIG{ALRM}='DEFAULT';
     $SIG{__DIE__}=\&catchexception;
@@ -753,6 +907,83 @@ sub logthis {
     print $fh "$local ($$) [$conserver] [$status]: $message\n";
 }
 
+#--------------------------------------  londtransaction:
+#  
+#  Performs a transaction with lond with timeout support.
+#    result = londtransaction(socket,request,timeout)
+#
+sub londtransaction {
+    my ($socket, $request, $tmo) = @_;
+
+    if($DEBUG) {
+	&logthis("londtransaction request: $request");
+    }
+
+    # Set the signal handlers: ALRM for timeout and disble the others.
+
+    $SIG{ALRM} = sub { die "timeout" };
+    $SIG{__DIE__} = 'DEFAULT';
+    
+    # Disable all but alarm so that only that can interupt the
+    # send /receive.
+    #
+    my $sigset = POSIX::SigSet->new(QUIT, USR1, HUP, INT, TERM);
+    my $priorsigs = POSIX::SigSet->new;
+    unless (defined sigprocmask(SIG_BLOCK, $sigset, $priorsigs)) {
+	&logthis("<font color=red> CRITICAL -- londtransaction ".
+		"failed to block signals </font>");
+	die "could not block signals in londtransaction";
+    }
+    $answer = '';
+    #
+    #  Send request to lond.
+    #
+    eval { 
+	alarm($tmo);
+	print $socket "$request\n";
+	alarm(0);
+    };
+    #  If request didn't timeout, try for the response.
+    #
+
+    if ($@!~/timeout/) {
+	eval {
+	    alarm($tmo);
+	    $answer = <$socket>;
+	    if($DEBUG) {
+		&logthis("Received $answer in londtransaction");
+	    }
+	    alarm(0);
+	};
+    } else {
+	if($DEBUG) {
+	    &logthis("Timeout on send in londtransaction");
+	}
+    }
+    if( ($@ =~ /timeout/)  && ($DEBUG)) {
+	&logthis("Timeout on receive in londtransaction");
+    }
+    #
+    # Restore the initial sigmask set.
+    #
+    unless (defined sigprocmask(SIG_UNBLOCK, $priorsigs)) {
+	&logthis("<font color=red> CRITICAL -- londtransaction ".
+		"failed to re-enable signal processing. </font>");
+	die "londtransaction failed to re-enable signals";
+    }
+    #
+    # go back to the prior handler set.
+    #
+    $SIG{ALRM} = 'DEFAULT';
+    $SIG{__DIE__} = \&cathcexception;
+
+    #    chomp $answer;
+    if ($DEBUG) {
+	&logthis("Returning $answer in londtransaction");
+    }
+    return $answer;
+
+}
 
 sub logperm {
     my $message=shift;
@@ -797,22 +1028,104 @@ lonc - LON TCP-MySQL-Server Daemon for h
 
 =head1 SYNOPSIS
 
+Usage: B<lonc>
+
 Should only be run as user=www.  This is a command-line script which
-is invoked by loncron.
+is invoked by B<loncron>.  There is no expectation that a typical user
+will manually start B<lonc> from the command-line.  (In other words,
+DO NOT START B<lonc> YOURSELF.)
 
 =head1 DESCRIPTION
 
 Provides persistent TCP connections to the other servers in the network
 through multiplexed domain sockets
 
- PID in subdir logs/lonc.pid
- kill kills
- HUP restarts
- USR1 tries to open connections again
+B<lonc> forks off children processes that correspond to the other servers
+in the network.  Management of these processes can be done at the
+parent process level or the child process level.
+
+  After forking off the children, B<lonc> the B<parent> 
+executes a main loop which simply waits for processes to exit.
+As a process exits, a new process managing a link to the same
+peer as the exiting process is created.  
+
+B<logs/lonc.log> is the location of log messages.
+
+The process management is now explained in terms of linux shell commands,
+subroutines internal to this code, and signal assignments:
+
+=over 4
+
+=item *
+
+PID is stored in B<logs/lonc.pid>
+
+This is the process id number of the parent B<lonc> process.
+
+=item *
+
+SIGTERM and SIGINT
+
+Parent signal assignment:
+ $SIG{INT}  = $SIG{TERM} = \&HUNTSMAN;
+
+Child signal assignment:
+ $SIG{INT}  = 'DEFAULT'; (and SIGTERM is DEFAULT also)
+(The child dies and a SIGALRM is sent to parent, awaking parent from slumber
+ to restart a new child.)
+
+Command-line invocations:
+ B<kill> B<-s> SIGTERM I<PID>
+ B<kill> B<-s> SIGINT I<PID>
+
+Subroutine B<HUNTSMAN>:
+ This is only invoked for the B<lonc> parent I<PID>.
+This kills all the children, and then the parent.
+The B<lonc.pid> file is cleared.
+
+=item *
+
+SIGHUP
+
+Current bug:
+ This signal can only be processed the first time
+on the parent process.  Subsequent SIGHUP signals
+have no effect.
+
+Parent signal assignment:
+ $SIG{HUP}  = \&HUPSMAN;
+
+Child signal assignment:
+ none (nothing happens)
+
+Command-line invocations:
+ B<kill> B<-s> SIGHUP I<PID>
+
+Subroutine B<HUPSMAN>:
+ This is only invoked for the B<lonc> parent I<PID>,
+This kills all the children, and then the parent.
+The B<lonc.pid> file is cleared.
+
+=item *
+
+SIGUSR1
+
+Parent signal assignment:
+ $SIG{USR1} = \&USRMAN;
+
+Child signal assignment:
+ $SIG{USR1}= \&logstatus;
+
+Command-line invocations:
+ B<kill> B<-s> SIGUSR1 I<PID>
+
+Subroutine B<USRMAN>:
+ When invoked for the B<lonc> parent I<PID>,
+SIGUSR1 is sent to all the children, and the status of
+each connection is logged.
 
-=head1 README
 
-Not yet written.
+=back
 
 =head1 PREREQUISITES