--- loncom/loncnew 2003/09/15 09:24:49 1.23 +++ loncom/loncnew 2003/12/11 23:18:37 1.34 @@ -2,7 +2,7 @@ # The LearningOnline Network with CAPA # lonc maintains the connections to remote computers # -# $Id: loncnew,v 1.23 2003/09/15 09:24:49 foxr Exp $ +# $Id: loncnew,v 1.34 2003/12/11 23:18:37 albertel Exp $ # # Copyright Michigan State University Board of Trustees # @@ -43,83 +43,8 @@ # - Detect transmission timeouts. # -# Change log: -# $Log: loncnew,v $ -# Revision 1.23 2003/09/15 09:24:49 foxr -# Add use strict and fix all the fallout from that. -# -# Revision 1.22 2003/09/02 10:34:47 foxr -# - Fix errors in host dead detection logic (too many cases where the -# retries left were not getting incremented or just not checked). -# - Added some additional status to the ps axuww display: -# o Remaining retries on a host. -# o >>> DEAD <<< indicator if I've given up on a host. -# - Tested the SIGHUP will reset the retries remaining count (thanks to -# the above status stuff, and get allow the loncnew to re-try again -# on the host (thanks to the log). -# -# Revision 1.21 2003/08/26 09:19:51 foxr -# How embarrassing... put in the SocketTimeout function in loncnew and forgot -# to actually hook it into the LondTransaction. Added this to MakeLondConnection -# where it belongs... hopefully transactions (not just connection attempts) will -# timeout more speedily than the socket errors will catch it. -# -# Revision 1.20 2003/08/25 18:48:11 albertel -# - fixing a forgotten ; -# -# Revision 1.19 2003/08/19 09:31:46 foxr -# Get socket directory from configuration rather than the old hard coded test -# way that I forgot to un-hard code. -# -# Revision 1.18 2003/08/06 09:52:29 foxr -# Also needed to remember to fail in-flight transactions if their sends fail. -# -# Revision 1.17 2003/08/03 00:44:31 foxr -# 1. Correct handling of connection failure: Assume it means the host is -# unreachable and fail all of the queued transactions. Note that the -# inflight transactions should fail on their own time due either to timeout -# or send/receive failures. -# 2. Correct handling of logs for forced death signals. Pull the signal -# from the event watcher. -# -# Revision 1.16 2003/07/29 02:33:05 foxr -# Add SIGINT processing to child processes to toggle annoying trace mode -# on/off.. will try to use this to isolate the compute boud process issue. -# -# Revision 1.15 2003/07/15 02:07:05 foxr -# Added code for lonc/lond transaction timeouts. Who knows if it works right. -# The intent is for a timeout to fail any transaction in progress and kill -# off the sockt that timed out. -# -# Revision 1.14 2003/07/03 02:10:18 foxr -# Get all of the signals to work correctly. -# -# Revision 1.13 2003/07/02 01:31:55 foxr -# Added kill -HUP logic (restart). -# -# Revision 1.11 2003/06/25 01:54:44 foxr -# Fix more problems with transaction failure. -# -# Revision 1.10 2003/06/24 02:46:04 foxr -# Put a limit on the number of times we'll retry a connection. -# Start getting the signal stuff put in as well...note that need to get signals -# going or else the client will permanently give up on dead servers. -# -# Revision 1.9 2003/06/13 02:38:43 foxr -# Add logging in 'expected format' -# -# Revision 1.8 2003/06/11 02:04:35 foxr -# Support delayed transactions... this is done uniformly by encapsulating -# transactions in an object ... a LondTransaction that is implemented by -# LondTransaction.pm -# -# Revision 1.7 2003/06/03 01:59:39 foxr -# complete coding to support deferred transactions. -# -# use strict; use lib "/home/httpd/lib/perl/"; -use lib "/home/foxr/newloncapa/types"; use Event qw(:DEFAULT ); use POSIX qw(:signal_h); use POSIX; @@ -141,12 +66,6 @@ use LONCAPA::HashIterator; # # Disable all signals we might receive from outside for now. # -#$SIG{QUIT} = IGNORE; -#$SIG{HUP} = IGNORE; -#$SIG{USR1} = IGNORE; -#$SIG{INT} = IGNORE; -#$SIG{CHLD} = IGNORE; -#$SIG{__DIE__} = IGNORE; # Read the httpd configuration file to get perl variables @@ -159,13 +78,15 @@ my %perlvar = %{$perlvarref}; # parent and shared variables. my %ChildHash; # by pid -> host. +my %HostToPid; # By host -> pid. +my %HostHash; # by loncapaname -> IP. my $MaxConnectionCount = 10; # Will get from config later. my $ClientConnection = 0; # Uniquifier for client events. my $DebugLevel = 0; -my $NextDebugLevel= 10; # So Sigint can toggle this. +my $NextDebugLevel= 2; # So Sigint can toggle this. my $IdleTimeout= 3600; # Wait an hour before pruning connections. # @@ -182,8 +103,8 @@ my $ConnectionCount = 0; my $IdleSeconds = 0; # Number of seconds idle. my $Status = ""; # Current status string. my $RecentLogEntry = ""; -my $ConnectionRetries=5; # Number of connection retries allowed. -my $ConnectionRetriesLeft=5; # Number of connection retries remaining. +my $ConnectionRetries=2; # Number of connection retries allowed. +my $ConnectionRetriesLeft=2; # Number of connection retries remaining. # # The hash below gives the HTML format for log messages @@ -384,6 +305,14 @@ sub Tick { my $Socket = $ActiveTransactions{$item}->getServer(); $Socket->Tick(); } + foreach my $item (keys %ActiveConnections) { + my $State = $ActiveConnections{$item}->data->GetState(); + if ($State ne 'Idle' && $State ne 'SendingRequest' && + $State ne 'ReceivingReply') { + Debug(5,"Ticking Socket $State $item"); + $ActiveConnections{$item}->data->Tick(); + } + } # Do we have work in the queue, but no connections to service them? # If so, try to make some new connections to get things going again. # @@ -393,18 +322,18 @@ sub Tick { if ($ConnectionRetriesLeft > 0) { my $Connections = ($Requests <= $MaxConnectionCount) ? $Requests : $MaxConnectionCount; - Debug(1,"Work but no connections, start ".$Connections." of them"); + Debug(5,"Work but no connections, start ".$Connections." of them"); my $successCount = 0; for (my $i =0; $i < $Connections; $i++) { $successCount += MakeLondConnection(); } if($successCount == 0) { # All connections failed: - Debug(1,"Work in queue failed to make any connectiouns\n"); + Debug(5,"Work in queue failed to make any connectiouns\n"); EmptyQueue(); # Fail pending transactions with con_lost. } } else { ShowStatus(GetServerHost()." >>> DEAD!!! <<<"); - Debug(1,"Work in queue, but gave up on connections..flushing\n"); + Debug(5,"Work in queue, but gave up on connections..flushing\n"); EmptyQueue(); # Connections can't be established. } @@ -429,7 +358,7 @@ Trigger disconnections of idle sockets. sub SetupTimer { Debug(6, "SetupTimer"); - Event->timer(interval => 1, debug => 1, cb => \&Tick ); + Event->timer(interval => 1, cb => \&Tick ); } =pod @@ -451,19 +380,19 @@ sub ServerToIdle { my $Socket = shift; # Get the socket. delete($ActiveTransactions{$Socket}); # Server has no transaction - &Debug(6, "Server to idle"); + &Debug(5, "Server to idle"); # If there's work to do, start the transaction: my $reqdata = $WorkQueue->dequeue(); # This is a LondTransaction - unless($reqdata eq undef) { - Debug(9, "Queue gave request data: ".$reqdata->getRequest()); + if ($reqdata ne undef) { + Debug(5, "Queue gave request data: ".$reqdata->getRequest()); &StartRequest($Socket, $reqdata); } else { # There's no work waiting, so push the server to idle list. - &Debug(8, "No new work requests, server connection going idle"); + &Debug(5, "No new work requests, server connection going idle"); $IdleConnections->push($Socket); } } @@ -509,7 +438,7 @@ sub ClientWritable { # request. &Debug(9,"Send result is ".$result." Defined: ".defined($result)); - if(defined($result)) { + if($result ne undef) { &Debug(9, "send result was defined"); if($result == length($Data)) { # Entire string sent. &Debug(9, "ClientWritable data all written"); @@ -580,7 +509,7 @@ The transaction that is being completed. =cut sub CompleteTransaction { - &Debug(6,"Complete transaction"); + &Debug(5,"Complete transaction"); my $Socket = shift; my $Transaction = shift; @@ -619,7 +548,6 @@ sub StartClientReply { &Debug(8," Reply was: ".$data); my $Serial = $ActiveClients{$Client}; my $desc = sprintf("Connection to lonc client %d", - $Serial); Event->io(fd => $Client, poll => "w", @@ -868,7 +796,7 @@ sub LondReadable { $ConnectionRetriesLeft = $ConnectionRetries; # success resets the count $Watcher->cancel(); if(exists($ActiveTransactions{$Socket})) { - Debug(8,"Completing transaction!!"); + Debug(5,"Completing transaction!!"); CompleteTransaction($Socket, $ActiveTransactions{$Socket}); } else { @@ -1115,7 +1043,7 @@ sub MakeLondConnection { my $Connection = LondConnection->new(&GetServerHost(), &GetServerPort()); - if($Connection == undef) { # Needs to be more robust later. + if($Connection eq undef) { # Needs to be more robust later. Log("CRITICAL","Failed to make a connection with lond."); $ConnectionRetriesLeft--; return 0; # Failure. @@ -1127,7 +1055,7 @@ sub MakeLondConnection { # exchange underway. # my $Socket = $Connection->GetSocket(); - if($Socket == undef) { + if($Socket eq undef) { die "did not get a socket from the connection"; } else { &Debug(9,"MakeLondConnection got socket: ".$Socket); @@ -1230,11 +1158,11 @@ sub QueueTransaction { my $LondSocket = $IdleConnections->pop(); if(!defined $LondSocket) { # Need to queue request. - Debug(8,"Must queue..."); + Debug(5,"Must queue..."); $WorkQueue->enqueue($requestData); if($ConnectionCount < $MaxConnectionCount) { if($ConnectionRetriesLeft > 0) { - Debug(4,"Starting additional lond connection"); + Debug(5,"Starting additional lond connection"); if(MakeLondConnection() == 0) { EmptyQueue(); # Fail transactions, can't make connection. } @@ -1272,7 +1200,7 @@ sub ClientRequest { my $rv = $socket->recv($thisread, POSIX::BUFSIZ, 0); Debug(8, "rcv: data length = ".length($thisread) ." read =".$thisread); - unless (defined $rv && length($thisread)) { + unless (defined $rv && length($thisread)) { # Likely eof on socket. Debug(5,"Client Socket closed on lonc for ".$RemoteHost); close($socket); @@ -1517,7 +1445,8 @@ sub CreateChild { my $pid = fork; if($pid) { # Parent $RemoteHost = "Parent"; - $ChildHash{$pid} = $RemoteHost; + $ChildHash{$pid} = $host; + $HostToPid{$host}= $pid; sigprocmask(SIG_UNBLOCK, $sigset); } else { # child. @@ -1575,11 +1504,13 @@ ShowStatus("Forking node servers"); Log("CRITICAL", "--------------- Starting children ---------------"); +LondConnection::ReadConfig; # Read standard config files. my $HostIterator = LondConnection::GetHostIterator; while (! $HostIterator->end()) { my $hostentryref = $HostIterator->get(); CreateChild($hostentryref->[0]); + $HostHash{$hostentryref->[0]} = $hostentryref->[4]; $HostIterator->next(); } $RemoteHost = "Parent Server"; @@ -1596,11 +1527,13 @@ $SIG{INT} = \&Terminate; $SIG{TERM} = \&Terminate; $SIG{HUP} = \&Restart; $SIG{USR1} = \&CheckKids; +$SIG{USR2} = \&UpdateKids; # LonManage update request. while(1) { my $deadchild = wait(); if(exists $ChildHash{$deadchild}) { # need to restart. my $deadhost = $ChildHash{$deadchild}; + delete($HostToPid{$deadhost}); delete($ChildHash{$deadchild}); Log("WARNING","Lost child pid= ".$deadchild. "Connected to host ".$deadhost); @@ -1639,6 +1572,114 @@ sub CheckKids { =pod +=head1 UpdateKids + +parent's SIGUSR2 handler. This handler: + +=item + +Rereads the hosts file. + +=item + +Kills off (via sigint) children for hosts that have disappeared. + +=item + +QUITs children for hosts that already exist (this just forces a status display +and resets the connection retry count for that host. + +=item + +Starts new children for hosts that have been added to the hosts.tab file since +the start of the master program and maintains them. + +=cut + +sub UpdateKids { + + Log("INFO", "Updating connections via SIGUSR2"); + + # Just in case we need to kill our own lonc, we wait a few seconds to + # give it a chance to receive and relay lond's response to the + # re-init command. + # + + sleep(2); # Wait a couple of seconds. + + my %hosts; # Indexed by loncapa hostname, value=ip. + + # Need to re-read the host table: + + + LondConnection::ReadConfig(); + my $I = LondConnection::GetHostIterator; + while (! $I->end()) { + my $item = $I->get(); + $hosts{$item->[0]} = $item->[4]; + $I->next(); + } + + # The logic below is written for clarity not for efficiency. + # Since I anticipate that this function is only rarely called, that's + # appropriate. There are certainly ways to combine the loops below, + # and anyone wishing to obscure the logic is welcome to go for it. + # Note that we don't re-direct sigchild. Instead we do what's needed + # to the data structures that keep track of children to ensure that + # when sigchild is honored, no new child is born. + # + + # For each existing child; if it's host doesn't exist, kill the child. + + foreach my $child (keys %ChildHash) { + my $oldhost = $ChildHash{$child}; + if (!(exists $hosts{$oldhost})) { + Log("CRITICAL", "Killing child for $oldhost host no longer exists"); + delete $ChildHash{$child}; + delete $HostToPid{$oldhost}; + kill 'QUIT' => $child; + } + } + # For each remaining existing child; if it's host's ip has changed, + # Restart the child on the new IP. + + foreach my $child (keys %ChildHash) { + my $oldhost = $ChildHash{$child}; + my $oldip = $HostHash{$oldhost}; + if ($hosts{$oldhost} ne $oldip) { + + # kill the old child. + + Log("CRITICAL", "Killing child for $oldhost host ip has changed..."); + delete $ChildHash{$child}; + delete $HostToPid{$oldhost}; + kill 'QUIT' => $child; + + # Do the book-keeping needed to start a new child on the + # new ip. + + $HostHash{$oldhost} = $hosts{$oldhost}; + CreateChild($oldhost); + } + } + # Finally, for each new host, not in the host hash, create a + # enter the host and create a new child. + # Force a status display of any existing process. + + foreach my $host (keys %hosts) { + if(!(exists $HostHash{$host})) { + Log("INFO", "New host $host discovered in hosts.tab..."); + $HostHash{$host} = $hosts{$host}; + CreateChild($host); + } else { + kill 'HUP' => $HostToPid{$host}; # status display. + } + } +} + + +=pod + =head1 Restart Signal handler for HUP... all children are killed and @@ -1652,7 +1693,7 @@ sub Restart { Log("CRITICAL", "Restarting"); my $execdir = $perlvar{'lonDaemons'}; unlink("$execdir/logs/lonc.pid"); - exec("$execdir/lonc"); + exec("$execdir/loncnew"); } =pod