loncom/loncnew - diff

Return to loncnew CVS log

Up to [LON-CAPA] / loncom

Diff for /loncom/loncnew between versions 1.21 and 1.27

version 1.21, 2003/08/26 09:19:51	version 1.27, 2003/10/07 11:23:03
Line 45	Line 45

# Change log:	# Change log:
# $Log$	# $Log$
	# Revision 1.27 2003/10/07 11:23:03 foxr
	# Installed and tested code to process reinit in parent server.
	#
	# Revision 1.26 2003/09/30 11:11:17 foxr
	# Add book-keeping hashes to support the re-init procedure.
	#
	# Revision 1.25 2003/09/23 11:22:14 foxr
	# Tested ability to receive sigusr2 This is now logged and must be
	# properly implemented as a re-read of hosts and re-init of appropriate
	# children.
	#
	# Revision 1.24 2003/09/16 09:46:42 foxr
	# Added skeletal infrastructure to support SIGUSR2 update hosts request.
	#
	# 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	# Revision 1.21 2003/08/26 09:19:51 foxr
# How embarrassing... put in the SocketTimeout function in loncnew and forgot	# How embarrassing... put in the SocketTimeout function in loncnew and forgot
# to actually hook it into the LondTransaction. Added this to MakeLondConnection	# to actually hook it into the LondTransaction. Added this to MakeLondConnection
Line 90	Line 117
# Revision 1.10 2003/06/24 02:46:04 foxr	# Revision 1.10 2003/06/24 02:46:04 foxr
# Put a limit on the number of times we'll retry a connection.	# 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	# Start getting the signal stuff put in as well...note that need to get signals
# going or else 6the client will permanently give up on dead servers.	# going or else the client will permanently give up on dead servers.
#	#
# Revision 1.9 2003/06/13 02:38:43 foxr	# Revision 1.9 2003/06/13 02:38:43 foxr
# Add logging in 'expected format'	# Add logging in 'expected format'
Line 104	Line 131
# complete coding to support deferred transactions.	# complete coding to support deferred transactions.
#	#
#	#
	use strict;
use lib "/home/httpd/lib/perl/";	use lib "/home/httpd/lib/perl/";
use lib "/home/foxr/newloncapa/types";	use lib "/home/foxr/newloncapa/types";
use Event qw(:DEFAULT );	use Event qw(:DEFAULT );
Line 128 use LONCAPA::HashIterator;	Line 155 use LONCAPA::HashIterator;
#	#
# Disable all signals we might receive from outside for now.	# 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	# Read the httpd configuration file to get perl variables
Line 146 my %perlvar = %{$perlvarref};	Line 167 my %perlvar = %{$perlvarref};
# parent and shared variables.	# parent and shared variables.

my %ChildHash; # by pid -> host.	my %ChildHash; # by pid -> host.
	my %HostToPid; # By host -> pid.
	my %HostHash; # by loncapaname -> IP.


my $MaxConnectionCount = 10; # Will get from config later.	my $MaxConnectionCount = 10; # Will get from config later.
Line 268 sub GetPeername {	Line 291 sub GetPeername {
my $peerip;	my $peerip;
if($AdrFamily == AF_INET) {	if($AdrFamily == AF_INET) {
($peerport, $peerip) = sockaddr_in($peer);	($peerport, $peerip) = sockaddr_in($peer);
my $peername = gethostbyaddr($iaddr, $AdrFamily);	my $peername = gethostbyaddr($peerip, $AdrFamily);
return $peername;	return $peername;
} elsif ($AdrFamily == AF_UNIX) {	} elsif ($AdrFamily == AF_UNIX) {
my $peerfile;	my $peerfile;
Line 289 sub Debug {	Line 312 sub Debug {
my $level = shift;	my $level = shift;
my $message = shift;	my $message = shift;
if ($level <= $DebugLevel) {	if ($level <= $DebugLevel) {
Log("INFO", "-Debug- $message host = $RemotHost");	Log("INFO", "-Debug- $message host = $RemoteHost");
}	}
}	}

Line 329 sub ShowStatus {	Line 352 sub ShowStatus {
sub SocketTimeout {	sub SocketTimeout {
my $Socket = shift;	my $Socket = shift;

KillSocket($Socket);	KillSocket($Socket); # A transaction timeout also counts as
	# a connection failure:
	$ConnectionRetriesLeft--;
}	}

=pod	=pod
Line 343 Invoked each timer tick.	Line 368 Invoked each timer tick.

sub Tick {	sub Tick {
my $client;	my $client;
ShowStatus(GetServerHost()." Connection count: ".$ConnectionCount);	if($ConnectionRetriesLeft > 0) {
	ShowStatus(GetServerHost()." Connection count: ".$ConnectionCount
	." Retries remaining: ".$ConnectionRetriesLeft);
	} else {
	ShowStatus(GetServerHost()." >> DEAD <<");
	}
# Is it time to prune connection count:	# Is it time to prune connection count:


Line 352 sub Tick {	Line 381 sub Tick {
($WorkQueue->Count() == 0)) { # Idle connections and nothing to do?	($WorkQueue->Count() == 0)) { # Idle connections and nothing to do?
$IdleSeconds++;	$IdleSeconds++;
if($IdleSeconds > $IdleTimeout) { # Prune a connection...	if($IdleSeconds > $IdleTimeout) { # Prune a connection...
$Socket = $IdleConnections->pop();	my $Socket = $IdleConnections->pop();
KillSocket($Socket);	KillSocket($Socket);
}	}
} else {	} else {
Line 361 sub Tick {	Line 390 sub Tick {
#	#
# For each inflight transaction, tick down its timeout counter.	# For each inflight transaction, tick down its timeout counter.
#	#
foreach $item (keys %ActiveTransactions) {	foreach my $item (keys %ActiveTransactions) {
my $Socket = $ActiveTransactions{$item}->getServer();	my $Socket = $ActiveTransactions{$item}->getServer();
$Socket->Tick();	$Socket->Tick();
}	}
Line 375 sub Tick {	Line 404 sub Tick {
my $Connections = ($Requests <= $MaxConnectionCount) ?	my $Connections = ($Requests <= $MaxConnectionCount) ?
$Requests : $MaxConnectionCount;	$Requests : $MaxConnectionCount;
Debug(1,"Work but no connections, start ".$Connections." of them");	Debug(1,"Work but no connections, start ".$Connections." of them");
for ($i =0; $i < $Connections; $i++) {	my $successCount = 0;
MakeLondConnection();	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");
	EmptyQueue(); # Fail pending transactions with con_lost.
}	}
} else {	} else {
	ShowStatus(GetServerHost()." >>> DEAD!!! <<<");
Debug(1,"Work in queue, but gave up on connections..flushing\n");	Debug(1,"Work in queue, but gave up on connections..flushing\n");
EmptyQueue(); # Connections can't be established.	EmptyQueue(); # Connections can't be established.
}	}
Line 430 sub ServerToIdle {	Line 465 sub ServerToIdle {

# If there's work to do, start the transaction:	# If there's work to do, start the transaction:

$reqdata = $WorkQueue->dequeue(); # This is a LondTransaction	my $reqdata = $WorkQueue->dequeue(); # This is a LondTransaction
unless($reqdata eq undef) {	unless($reqdata eq undef) {
Debug(9, "Queue gave request data: ".$reqdata->getRequest());	Debug(9, "Queue gave request data: ".$reqdata->getRequest());
&StartRequest($Socket, $reqdata);	&StartRequest($Socket, $reqdata);
Line 564 sub CompleteTransaction {	Line 599 sub CompleteTransaction {
StartClientReply($Transaction, $data);	StartClientReply($Transaction, $data);
} else { # Delete deferred transaction file.	} else { # Delete deferred transaction file.
Log("SUCCESS", "A delayed transaction was completed");	Log("SUCCESS", "A delayed transaction was completed");
LogPerm("S:$Client:".$Transaction->getRequest());	LogPerm("S:$Transaction->getClient() :".$Transaction->getRequest());
unlink $Transaction->getFile();	unlink $Transaction->getFile();
}	}
}	}
Line 632 sub FailTransaction {	Line 667 sub FailTransaction {
Debug(1," Replying con_lost to ".$transaction->getRequest());	Debug(1," Replying con_lost to ".$transaction->getRequest());
StartClientReply($transaction, "con_lost\n");	StartClientReply($transaction, "con_lost\n");
}	}
	if($ConnectionRetriesLeft <= 0) {
	Log("CRITICAL", "Host marked dead: ".GetServerHost());
	}

}	}

Line 643 sub FailTransaction {	Line 681 sub FailTransaction {

=cut	=cut
sub EmptyQueue {	sub EmptyQueue {
	$ConnectionRetriesLeft--; # Counts as connection failure too.
while($WorkQueue->Count()) {	while($WorkQueue->Count()) {
my $request = $WorkQueue->dequeue(); # This is a transaction	my $request = $WorkQueue->dequeue(); # This is a transaction
FailTransaction($request);	FailTransaction($request);
Line 657 Close all connections open on lond prior	Line 696 Close all connections open on lond prior

=cut	=cut
sub CloseAllLondConnections {	sub CloseAllLondConnections {
foreach $Socket (keys %ActiveConnections) {	foreach my $Socket (keys %ActiveConnections) {
KillSocket($Socket);	KillSocket($Socket);
}	}
}	}
Line 709 sub KillSocket {	Line 748 sub KillSocket {
# work queue, the work all gets failed with con_lost.	# work queue, the work all gets failed with con_lost.
#	#
if($ConnectionCount == 0) {	if($ConnectionCount == 0) {
EmptyQueue;	EmptyQueue();
}	}
}	}

Line 778 sub LondReadable {	Line 817 sub LondReadable {
my $Socket = $Watcher->data;	my $Socket = $Watcher->data;
my $client = undef;	my $client = undef;

&Debug(6,"LondReadable called state = ".$State);	&Debug(6,"LondReadable called state = ".$Socket->GetState());


my $State = $Socket->GetState(); # All action depends on the state.	my $State = $Socket->GetState(); # All action depends on the state.
Line 799 sub LondReadable {	Line 838 sub LondReadable {
}	}
$Watcher->cancel();	$Watcher->cancel();
KillSocket($Socket);	KillSocket($Socket);
	$ConnectionRetriesLeft--; # Counts as connection failure
return;	return;
}	}
SocketDump(6,$Socket);	SocketDump(6,$Socket);
Line 832 sub LondReadable {	Line 872 sub LondReadable {
} elsif ($State eq "Idle") {	} elsif ($State eq "Idle") {
# If necessary, complete a transaction and then go into the	# If necessary, complete a transaction and then go into the
# idle queue.	# idle queue.
	# Note that a trasition to idle indicates a live lond
	# on the other end so reset the connection retries.
	#
	$ConnectionRetriesLeft = $ConnectionRetries; # success resets the count
$Watcher->cancel();	$Watcher->cancel();
if(exists($ActiveTransactions{$Socket})) {	if(exists($ActiveTransactions{$Socket})) {
Debug(8,"Completing transaction!!");	Debug(8,"Completing transaction!!");
Line 1044 sub QueueDelayed {	Line 1088 sub QueueDelayed {
Debug(4, "Delayed path: ".$path);	Debug(4, "Delayed path: ".$path);
opendir(DIRHANDLE, $path);	opendir(DIRHANDLE, $path);

@alldelayed = grep /\.$RemoteHost$/, readdir DIRHANDLE;	my @alldelayed = grep /\.$RemoteHost$/, readdir DIRHANDLE;
Debug(4, "Got ".$alldelayed." delayed files");
closedir(DIRHANDLE);	closedir(DIRHANDLE);
my $dfname;	my $dfname;
my $reqfile;	my $reqfile;
Line 1087 sub MakeLondConnection {	Line 1130 sub MakeLondConnection {
$ConnectionRetriesLeft--;	$ConnectionRetriesLeft--;
return 0; # Failure.	return 0; # Failure.
} else {	} else {
$ConnectionRetriesLeft = $ConnectionRetries; # success resets the count
# The connection needs to have writability	# The connection needs to have writability
# monitored in order to send the init sequence	# monitored in order to send the init sequence
# that starts the whole authentication/key	# that starts the whole authentication/key
Line 1102 sub MakeLondConnection {	Line 1145 sub MakeLondConnection {

$Connection->SetTimeoutCallback(\&SocketTimeout);	$Connection->SetTimeoutCallback(\&SocketTimeout);

$event = Event->io(fd => $Socket,	my $event = Event->io(fd => $Socket,
poll => 'w',	poll => 'w',
cb => \&LondWritable,	cb => \&LondWritable,
data => $Connection,	data => $Connection,
Line 1158 sub StartRequest {	Line 1201 sub StartRequest {
$ActiveTransactions{$Lond} = $Request;	$ActiveTransactions{$Lond} = $Request;

$Lond->InitiateTransaction($Request->getRequest());	$Lond->InitiateTransaction($Request->getRequest());
$event = Event->io(fd => $Socket,	my $event = Event->io(fd => $Socket,
poll => "w",	poll => "w",
cb => \&LondWritable,	cb => \&LondWritable,
data => $Lond,	data => $Lond,
Line 1200 sub QueueTransaction {	Line 1243 sub QueueTransaction {
Debug(8,"Must queue...");	Debug(8,"Must queue...");
$WorkQueue->enqueue($requestData);	$WorkQueue->enqueue($requestData);
if($ConnectionCount < $MaxConnectionCount) {	if($ConnectionCount < $MaxConnectionCount) {
Debug(4,"Starting additional lond connection");	if($ConnectionRetriesLeft > 0) {
if(MakeLondConnection() == 0) {	Debug(4,"Starting additional lond connection");
EmptyQueue(); # Fail transactions, can't make connection.	if(MakeLondConnection() == 0) {
	EmptyQueue(); # Fail transactions, can't make connection.
	}
	} else {
	ShowStatus(GetServerHost()." >>> DEAD !!!! <<<");
	EmptyQueue(); # It's worse than that ... he's dead Jim.
}	}
}	}
} else { # Can start the request:	} else { # Can start the request:
Line 1368 sub SetupLoncListener {	Line 1416 sub SetupLoncListener {
Child USR1 signal handler to report the most recent status	Child USR1 signal handler to report the most recent status
into the status file.	into the status file.

	We also use this to reset the retries count in order to allow the
	client to retry connections with a previously dead server.
=cut	=cut
sub ChildStatus {	sub ChildStatus {
my $event = shift;	my $event = shift;
Line 1378 sub ChildStatus {	Line 1428 sub ChildStatus {
my $fh = IO::File->new(">>$docdir/lon-status/loncstatus.txt");	my $fh = IO::File->new(">>$docdir/lon-status/loncstatus.txt");
print $fh $$."\t".$RemoteHost."\t".$Status."\t".	print $fh $$."\t".$RemoteHost."\t".$Status."\t".
$RecentLogEntry."\n";	$RecentLogEntry."\n";
	$ConnectionRetriesLeft = $ConnectionRetries;
}	}

=pod	=pod
Line 1473 sub CreateChild {	Line 1524 sub CreateChild {
my $host = shift;	my $host = shift;
$RemoteHost = $host;	$RemoteHost = $host;
Log("CRITICAL", "Forking server for ".$host);	Log("CRITICAL", "Forking server for ".$host);
$pid = fork;	my $pid = fork;
if($pid) { # Parent	if($pid) { # Parent
$RemoteHost = "Parent";	$RemoteHost = "Parent";
$ChildHash{$pid} = $RemoteHost;	$ChildHash{$pid} = $host;
	$HostToPid{$host}= $pid;
sigprocmask(SIG_UNBLOCK, $sigset);	sigprocmask(SIG_UNBLOCK, $sigset);

} else { # child.	} else { # child.
ShowStatus("Connected to ".$RemoteHost);	ShowStatus("Connected to ".$RemoteHost);
$SIG{INT} = DEFAULT;	$SIG{INT} = 'DEFAULT';
sigprocmask(SIG_UNBLOCK, $sigset);	sigprocmask(SIG_UNBLOCK, $sigset);
ChildProcess; # Does not return.	ChildProcess; # Does not return.
}	}
Line 1518 if ($childpid != 0) {	Line 1570 if ($childpid != 0) {
#	#

ShowStatus("Parent writing pid file:");	ShowStatus("Parent writing pid file:");
$execdir = $perlvar{'lonDaemons'};	my $execdir = $perlvar{'lonDaemons'};
open (PIDSAVE, ">$execdir/logs/lonc.pid");	open (PIDSAVE, ">$execdir/logs/lonc.pid");
print PIDSAVE "$$\n";	print PIDSAVE "$$\n";
close(PIDSAVE);	close(PIDSAVE);
Line 1537 Log("CRITICAL", "--------------- Startin	Line 1589 Log("CRITICAL", "--------------- Startin
my $HostIterator = LondConnection::GetHostIterator;	my $HostIterator = LondConnection::GetHostIterator;
while (! $HostIterator->end()) {	while (! $HostIterator->end()) {

$hostentryref = $HostIterator->get();	my $hostentryref = $HostIterator->get();
CreateChild($hostentryref->[0]);	CreateChild($hostentryref->[0]);
	$HostHash{$hostentryref->[0]} = $hostentryref->[4];
$HostIterator->next();	$HostIterator->next();
}	}
$RemoteHost = "Parent Server";	$RemoteHost = "Parent Server";
Line 1555 $SIG{INT} = \&Terminate;	Line 1608 $SIG{INT} = \&Terminate;
$SIG{TERM} = \&Terminate;	$SIG{TERM} = \&Terminate;
$SIG{HUP} = \&Restart;	$SIG{HUP} = \&Restart;
$SIG{USR1} = \&CheckKids;	$SIG{USR1} = \&CheckKids;
	$SIG{USR2} = \&UpdateKids; # LonManage update request.

while(1) {	while(1) {
$deadchild = wait();	my $deadchild = wait();
if(exists $ChildHash{$deadchild}) { # need to restart.	if(exists $ChildHash{$deadchild}) { # need to restart.
$deadhost = $ChildHash{$deadchild};	my $deadhost = $ChildHash{$deadchild};
	delete($HostToPid{$deadhost});
delete($ChildHash{$deadchild});	delete($ChildHash{$deadchild});
Log("WARNING","Lost child pid= ".$deadchild.	Log("WARNING","Lost child pid= ".$deadchild.
"Connected to host ".$deadhost);	"Connected to host ".$deadhost);
Line 1589 sub CheckKids {	Line 1644 sub CheckKids {
my $now=time;	my $now=time;
my $local=localtime($now);	my $local=localtime($now);
print $fh "LONC status $local - parent $$ \n\n";	print $fh "LONC status $local - parent $$ \n\n";
foreach $pid (keys %ChildHash) {	foreach my $pid (keys %ChildHash) {
Debug(2, "Sending USR1 -> $pid");	Debug(2, "Sending USR1 -> $pid");
kill 'USR1' => $pid; # Tell Child to report status.	kill 'USR1' => $pid; # Tell Child to report status.
sleep 1; # Wait so file doesn't intermix.	sleep 1; # Wait so file doesn't intermix.
Line 1598 sub CheckKids {	Line 1653 sub CheckKids {

=pod	=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	=head1 Restart

Signal handler for HUP... all children are killed and	Signal handler for HUP... all children are killed and
Line 1607 the config file.	Line 1770 the config file.
=cut	=cut

sub Restart {	sub Restart {
KillThemAll; # First kill all the children.	&KillThemAll; # First kill all the children.
Log("CRITICAL", "Restarting");	Log("CRITICAL", "Restarting");
my $execdir = $perlvar{'lonDaemons'};	my $execdir = $perlvar{'lonDaemons'};
unlink("$execdir/logs/lonc.pid");	unlink("$execdir/logs/lonc.pid");
Line 1626 SIGHUP. Responds to sigint and sigterm.	Line 1789 SIGHUP. Responds to sigint and sigterm.
sub KillThemAll {	sub KillThemAll {
Debug(2, "Kill them all!!");	Debug(2, "Kill them all!!");
local($SIG{CHLD}) = 'IGNORE'; # Our children >will< die.	local($SIG{CHLD}) = 'IGNORE'; # Our children >will< die.
foreach $pid (keys %ChildHash) {	foreach my $pid (keys %ChildHash) {
my $serving = $ChildHash{$pid};	my $serving = $ChildHash{$pid};
Debug(2, "Killing lonc for $serving pid = $pid");	Debug(2, "Killing lonc for $serving pid = $pid");
ShowStatus("Killing lonc for $serving pid = $pid");	ShowStatus("Killing lonc for $serving pid = $pid");

FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>

Removed from v.1.21
changed lines
	Added in v.1.27