--- loncom/loncnew 2004/06/01 10:02:13 1.47 +++ loncom/loncnew 2004/08/30 11:01:04 1.52 @@ -2,7 +2,7 @@ # The LearningOnline Network with CAPA # lonc maintains the connections to remote computers # -# $Id: loncnew,v 1.47 2004/06/01 10:02:13 foxr Exp $ +# $Id: loncnew,v 1.52 2004/08/30 11:01:04 foxr Exp $ # # Copyright Michigan State University Board of Trustees # @@ -82,7 +82,7 @@ my $ClientConnection = 0; # Uniquifier f my $DebugLevel = 0; my $NextDebugLevel= 2; # So Sigint can toggle this. -my $IdleTimeout= 3600; # Wait an hour before pruning connections. +my $IdleTimeout= 600; # Wait 10 minutes before pruning connections. my $LogTransactions = 0; # When True, all transactions/replies get logged. @@ -103,6 +103,7 @@ my $RecentLogEntry = ""; my $ConnectionRetries=2; # Number of connection retries allowed. my $ConnectionRetriesLeft=2; # Number of connection retries remaining. my $LondVersion = "unknown"; # Version of lond we talk with. +my $KeyMode = ""; # e.g. ssl, local, insecure from last connect. # # The hash below gives the HTML format for log messages @@ -232,7 +233,7 @@ sub SocketDump { my ($level, $socket) = @_; if($level <= $DebugLevel) { - $socket->Dump(); + $socket->Dump(-1); # Ensure it will get dumped. } } @@ -264,8 +265,8 @@ sub ShowStatus { sub SocketTimeout { my $Socket = shift; Log("WARNING", "A socket timeout was detected"); - Debug(0, " SocketTimeout called: "); - $Socket->Dump(); + Debug(5, " SocketTimeout called: "); + $Socket->Dump(0); if(exists($ActiveTransactions{$Socket})) { FailTransaction($ActiveTransactions{$Socket}); } @@ -273,7 +274,7 @@ sub SocketTimeout { # a connection failure: $ConnectionRetriesLeft--; if($ConnectionRetriesLeft <= 0) { - Log("CRITICAL", "Host marked dead: ".GetServerHost()); + Log("CRITICAL", "Host marked DEAD: ".GetServerHost()); } } @@ -289,10 +290,14 @@ Invoked each timer tick. sub Tick { + my ($Event) = @_; + my $clock_watcher = $Event->w; + my $client; if($ConnectionRetriesLeft > 0) { ShowStatus(GetServerHost()." Connection count: ".$ConnectionCount - ." Retries remaining: ".$ConnectionRetriesLeft); + ." Retries remaining: ".$ConnectionRetriesLeft + ." ($KeyMode)"); } else { ShowStatus(GetServerHost()." >> DEAD <<"); } @@ -347,6 +352,10 @@ sub Tick { } } + if ($ConnectionCount == 0) { + $KeyMode = ""; + $clock_watcher->cancel(); + } } =pod @@ -387,6 +396,7 @@ long enough, it will be shut down and re sub ServerToIdle { my $Socket = shift; # Get the socket. + $KeyMode = $Socket->{AuthenticationMode}; delete($ActiveTransactions{$Socket}); # Server has no transaction &Debug(5, "Server to idle"); @@ -466,7 +476,7 @@ sub ClientWritable { } else { # Partial string sent. $Watcher->data(substr($Data, $result)); if($result == 0) { # client hung up on us!! - Log("INFO", "lonc pipe client hung up on us!"); + # Log("INFO", "lonc pipe client hung up on us!"); $Watcher->cancel; $Socket->shutdown(2); $Socket->close(); @@ -594,7 +604,13 @@ Parameters: sub FailTransaction { my $transaction = shift; - Log("WARNING", "Failing transaction ".$transaction->getRequest()); + + # If the socket is dead, that's already logged. + + if ($ConnectionRetriesLeft > 0) { + Log("WARNING", "Failing transaction " + .$transaction->getRequest()); + } Debug(1, "Failing transaction: ".$transaction->getRequest()); if (!$transaction->isDeferred()) { # If the transaction is deferred we'll get to it. my $client = $transaction->getClient(); @@ -1128,7 +1144,9 @@ sub MakeLondConnection { data => $Connection, desc => 'Connection to lond server'); $ActiveConnections{$Connection} = $event; - + if ($ConnectionCount == 0) { + &SetupTimer; # Need to handle timeouts with connections... + } $ConnectionCount++; Debug(4, "Connection count = ".$ConnectionCount); if($ConnectionCount == 1) { # First Connection: @@ -1441,7 +1459,7 @@ sub ChildStatus { my $state = $Socket->GetState(); print $fh "Connection $i State: $state\n"; print STDERR "---------------------- Connection $i \n"; - $Socket->Dump(); + $Socket->Dump(-1); # Ensure it gets dumped.. $i++; } } @@ -1515,7 +1533,6 @@ sub ChildProcess { cb => \&ToggleDebug, data => "INT"); - SetupTimer(); SetupLoncListener(); @@ -1538,9 +1555,10 @@ sub ChildProcess { # Create a new child for host passed in: sub CreateChild { + my $host = shift; + my $sigset = POSIX::SigSet->new(SIGINT); sigprocmask(SIG_BLOCK, $sigset); - my $host = shift; $RemoteHost = $host; Log("CRITICAL", "Forking server for ".$host); my $pid = fork; @@ -1811,17 +1829,33 @@ sub KillThemAll { local($SIG{CHLD}) = 'IGNORE'; # Our children >will< die. foreach my $pid (keys %ChildHash) { my $serving = $ChildHash{$pid}; - Debug(2, "Killing lonc for $serving pid = $pid"); - ShowStatus("Killing lonc for $serving pid = $pid"); - Log("CRITICAL", "Killing lonc for $serving pid = $pid"); + ShowStatus("Nicely Killing lonc for $serving pid = $pid"); + Log("CRITICAL", "Nicely Killing lonc for $serving pid = $pid"); kill 'QUIT' => $pid; - delete($ChildHash{$pid}); } - my $execdir = $perlvar{'lonDaemons'}; - unlink("$execdir/logs/lonc.pid"); + } + +# +# Kill all children via KILL. Just in case the +# first shot didn't get them. + +sub really_kill_them_all_dammit +{ + Debug(2, "Kill them all Dammit"); + local($SIG{CHLD} = 'IGNORE'); # In case some purist reenabled them. + foreach my $pid (keys %ChildHash) { + my $serving = $ChildHash{$pid}; + &ShowStatus("Nastily killing lonc for $serving pid = $pid"); + Log("CRITICAL", "Nastily killing lonc for $serving pid = $pid"); + kill 'KILL' => $pid; + delete($ChildHash{$pid}); + my $execdir = $perlvar{'lonDaemons'}; + unlink("$execdir/logs/lonc.pid"); + } +} =pod =head1 Terminate @@ -1831,7 +1865,15 @@ Terminate the system. =cut sub Terminate { - KillThemAll; + &Log("CRITICAL", "Asked to kill children.. first be nice..."); + &KillThemAll; + # + # By now they really should all be dead.. but just in case + # send them all SIGKILL's after a bit of waiting: + + sleep(4); + &Log("CRITICAL", "Now kill children nasty"); + &really_kill_them_all_dammit; Log("CRITICAL","Master process exiting"); exit 0;