|
version 1.24, 2003/09/16 09:46:42
|
version 1.35, 2003/12/12 09:57:28
|
|
Line 43
|
Line 43
|
| # - Detect transmission timeouts. |
# - Detect transmission timeouts. |
| # |
# |
| |
|
| # Change log: |
|
| # $Log$ |
|
| # 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 |
|
| # 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 strict; |
| use lib "/home/httpd/lib/perl/"; |
use lib "/home/httpd/lib/perl/"; |
| use lib "/home/foxr/newloncapa/types"; |
|
| use Event qw(:DEFAULT ); |
use Event qw(:DEFAULT ); |
| use POSIX qw(:signal_h); |
use POSIX qw(:signal_h); |
| use POSIX; |
use POSIX; |
|
Line 144 use LONCAPA::HashIterator;
|
Line 66 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 162 my %perlvar = %{$perlvarref};
|
Line 78 my %perlvar = %{$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. |
| my $ClientConnection = 0; # Uniquifier for client events. |
my $ClientConnection = 0; # Uniquifier for client events. |
| |
|
| my $DebugLevel = 0; |
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. |
my $IdleTimeout= 3600; # Wait an hour before pruning connections. |
| |
|
| # |
# |
|
Line 185 my $ConnectionCount = 0;
|
Line 103 my $ConnectionCount = 0;
|
| my $IdleSeconds = 0; # Number of seconds idle. |
my $IdleSeconds = 0; # Number of seconds idle. |
| my $Status = ""; # Current status string. |
my $Status = ""; # Current status string. |
| my $RecentLogEntry = ""; |
my $RecentLogEntry = ""; |
| my $ConnectionRetries=5; # Number of connection retries allowed. |
my $ConnectionRetries=2; # Number of connection retries allowed. |
| my $ConnectionRetriesLeft=5; # Number of connection retries remaining. |
my $ConnectionRetriesLeft=2; # Number of connection retries remaining. |
| |
|
| # |
# |
| # The hash below gives the HTML format for log messages |
# The hash below gives the HTML format for log messages |
|
Line 292 sub GetPeername {
|
Line 210 sub GetPeername {
|
| return $peerfile; |
return $peerfile; |
| } |
} |
| } |
} |
| #----------------------------- Timer management ------------------------ |
|
| =pod |
=pod |
| |
|
| =head2 Debug |
=head2 Debug |
|
Line 349 sub SocketTimeout {
|
Line 266 sub SocketTimeout {
|
| # a connection failure: |
# a connection failure: |
| $ConnectionRetriesLeft--; |
$ConnectionRetriesLeft--; |
| } |
} |
| |
#----------------------------- Timer management ------------------------ |
| |
|
| =pod |
=pod |
| |
|
|
Line 383 sub Tick {
|
Line 301 sub Tick {
|
| # |
# |
| # For each inflight transaction, tick down its timeout counter. |
# For each inflight transaction, tick down its timeout counter. |
| # |
# |
| foreach my $item (keys %ActiveTransactions) { |
|
| my $Socket = $ActiveTransactions{$item}->getServer(); |
foreach my $item (keys %ActiveConnections) { |
| $Socket->Tick(); |
my $State = $ActiveConnections{$item}->data->GetState(); |
| |
if ($State ne 'Idle') { |
| |
Debug(5,"Ticking Socket $State $item"); |
| |
$ActiveConnections{$item}->data->Tick(); |
| |
} |
| } |
} |
| # Do we have work in the queue, but no connections to service them? |
# 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. |
# If so, try to make some new connections to get things going again. |
|
Line 396 sub Tick {
|
Line 318 sub Tick {
|
| if ($ConnectionRetriesLeft > 0) { |
if ($ConnectionRetriesLeft > 0) { |
| my $Connections = ($Requests <= $MaxConnectionCount) ? |
my $Connections = ($Requests <= $MaxConnectionCount) ? |
| $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; |
my $successCount = 0; |
| for (my $i =0; $i < $Connections; $i++) { |
for (my $i =0; $i < $Connections; $i++) { |
| $successCount += MakeLondConnection(); |
$successCount += MakeLondConnection(); |
| } |
} |
| if($successCount == 0) { # All connections failed: |
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. |
EmptyQueue(); # Fail pending transactions with con_lost. |
| } |
} |
| } else { |
} else { |
| ShowStatus(GetServerHost()." >>> DEAD!!! <<<"); |
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. |
EmptyQueue(); # Connections can't be established. |
| } |
} |
| |
|
|
Line 432 Trigger disconnections of idle sockets.
|
Line 354 Trigger disconnections of idle sockets.
|
| |
|
| sub SetupTimer { |
sub SetupTimer { |
| Debug(6, "SetupTimer"); |
Debug(6, "SetupTimer"); |
| Event->timer(interval => 1, debug => 1, cb => \&Tick ); |
Event->timer(interval => 1, cb => \&Tick ); |
| } |
} |
| |
|
| =pod |
=pod |
|
Line 454 sub ServerToIdle {
|
Line 376 sub ServerToIdle {
|
| my $Socket = shift; # Get the socket. |
my $Socket = shift; # Get the socket. |
| delete($ActiveTransactions{$Socket}); # Server has no transaction |
delete($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: |
# If there's work to do, start the transaction: |
| |
|
| my $reqdata = $WorkQueue->dequeue(); # This is a LondTransaction |
my $reqdata = $WorkQueue->dequeue(); # This is a LondTransaction |
| unless($reqdata eq undef) { |
if ($reqdata ne undef) { |
| Debug(9, "Queue gave request data: ".$reqdata->getRequest()); |
Debug(5, "Queue gave request data: ".$reqdata->getRequest()); |
| &StartRequest($Socket, $reqdata); |
&StartRequest($Socket, $reqdata); |
| |
|
| } else { |
} else { |
| |
|
| # There's no work waiting, so push the server to idle list. |
# 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); |
$IdleConnections->push($Socket); |
| } |
} |
| } |
} |
|
Line 512 sub ClientWritable {
|
Line 434 sub ClientWritable {
|
| # request. |
# request. |
| |
|
| &Debug(9,"Send result is ".$result." Defined: ".defined($result)); |
&Debug(9,"Send result is ".$result." Defined: ".defined($result)); |
| if(defined($result)) { |
if($result ne undef) { |
| &Debug(9, "send result was defined"); |
&Debug(9, "send result was defined"); |
| if($result == length($Data)) { # Entire string sent. |
if($result == length($Data)) { # Entire string sent. |
| &Debug(9, "ClientWritable data all written"); |
&Debug(9, "ClientWritable data all written"); |
|
Line 583 The transaction that is being completed.
|
Line 505 The transaction that is being completed.
|
| =cut |
=cut |
| |
|
| sub CompleteTransaction { |
sub CompleteTransaction { |
| &Debug(6,"Complete transaction"); |
&Debug(5,"Complete transaction"); |
| my $Socket = shift; |
my $Socket = shift; |
| my $Transaction = shift; |
my $Transaction = shift; |
| |
|
|
Line 622 sub StartClientReply {
|
Line 544 sub StartClientReply {
|
| &Debug(8," Reply was: ".$data); |
&Debug(8," Reply was: ".$data); |
| my $Serial = $ActiveClients{$Client}; |
my $Serial = $ActiveClients{$Client}; |
| my $desc = sprintf("Connection to lonc client %d", |
my $desc = sprintf("Connection to lonc client %d", |
| |
|
| $Serial); |
$Serial); |
| Event->io(fd => $Client, |
Event->io(fd => $Client, |
| poll => "w", |
poll => "w", |
|
Line 871 sub LondReadable {
|
Line 792 sub LondReadable {
|
| $ConnectionRetriesLeft = $ConnectionRetries; # success resets the count |
$ConnectionRetriesLeft = $ConnectionRetries; # success resets the count |
| $Watcher->cancel(); |
$Watcher->cancel(); |
| if(exists($ActiveTransactions{$Socket})) { |
if(exists($ActiveTransactions{$Socket})) { |
| Debug(8,"Completing transaction!!"); |
Debug(5,"Completing transaction!!"); |
| CompleteTransaction($Socket, |
CompleteTransaction($Socket, |
| $ActiveTransactions{$Socket}); |
$ActiveTransactions{$Socket}); |
| } else { |
} else { |
|
Line 1118 sub MakeLondConnection {
|
Line 1039 sub MakeLondConnection {
|
| my $Connection = LondConnection->new(&GetServerHost(), |
my $Connection = LondConnection->new(&GetServerHost(), |
| &GetServerPort()); |
&GetServerPort()); |
| |
|
| if($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."); |
Log("CRITICAL","Failed to make a connection with lond."); |
| $ConnectionRetriesLeft--; |
$ConnectionRetriesLeft--; |
| return 0; # Failure. |
return 0; # Failure. |
|
Line 1130 sub MakeLondConnection {
|
Line 1051 sub MakeLondConnection {
|
| # exchange underway. |
# exchange underway. |
| # |
# |
| my $Socket = $Connection->GetSocket(); |
my $Socket = $Connection->GetSocket(); |
| if($Socket == undef) { |
if($Socket eq undef) { |
| die "did not get a socket from the connection"; |
die "did not get a socket from the connection"; |
| } else { |
} else { |
| &Debug(9,"MakeLondConnection got socket: ".$Socket); |
&Debug(9,"MakeLondConnection got socket: ".$Socket); |
|
Line 1233 sub QueueTransaction {
|
Line 1154 sub QueueTransaction {
|
| |
|
| my $LondSocket = $IdleConnections->pop(); |
my $LondSocket = $IdleConnections->pop(); |
| if(!defined $LondSocket) { # Need to queue request. |
if(!defined $LondSocket) { # Need to queue request. |
| Debug(8,"Must queue..."); |
Debug(5,"Must queue..."); |
| $WorkQueue->enqueue($requestData); |
$WorkQueue->enqueue($requestData); |
| if($ConnectionCount < $MaxConnectionCount) { |
if($ConnectionCount < $MaxConnectionCount) { |
| if($ConnectionRetriesLeft > 0) { |
if($ConnectionRetriesLeft > 0) { |
| Debug(4,"Starting additional lond connection"); |
Debug(5,"Starting additional lond connection"); |
| if(MakeLondConnection() == 0) { |
if(MakeLondConnection() == 0) { |
| EmptyQueue(); # Fail transactions, can't make connection. |
EmptyQueue(); # Fail transactions, can't make connection. |
| } |
} |
|
Line 1275 sub ClientRequest {
|
Line 1196 sub ClientRequest {
|
| my $rv = $socket->recv($thisread, POSIX::BUFSIZ, 0); |
my $rv = $socket->recv($thisread, POSIX::BUFSIZ, 0); |
| Debug(8, "rcv: data length = ".length($thisread) |
Debug(8, "rcv: data length = ".length($thisread) |
| ." read =".$thisread); |
." read =".$thisread); |
| unless (defined $rv && length($thisread)) { |
unless (defined $rv && length($thisread)) { |
| # Likely eof on socket. |
# Likely eof on socket. |
| Debug(5,"Client Socket closed on lonc for ".$RemoteHost); |
Debug(5,"Client Socket closed on lonc for ".$RemoteHost); |
| close($socket); |
close($socket); |
|
Line 1520 sub CreateChild {
|
Line 1441 sub CreateChild {
|
| my $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. |
|
Line 1578 ShowStatus("Forking node servers");
|
Line 1500 ShowStatus("Forking node servers");
|
| |
|
| Log("CRITICAL", "--------------- Starting children ---------------"); |
Log("CRITICAL", "--------------- Starting children ---------------"); |
| |
|
| |
LondConnection::ReadConfig; # Read standard config files. |
| my $HostIterator = LondConnection::GetHostIterator; |
my $HostIterator = LondConnection::GetHostIterator; |
| while (! $HostIterator->end()) { |
while (! $HostIterator->end()) { |
| |
|
| my $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 1605 while(1) {
|
Line 1529 while(1) {
|
| my $deadchild = wait(); |
my $deadchild = wait(); |
| if(exists $ChildHash{$deadchild}) { # need to restart. |
if(exists $ChildHash{$deadchild}) { # need to restart. |
| my $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 1645 sub CheckKids {
|
Line 1570 sub CheckKids {
|
| |
|
| =head1 UpdateKids |
=head1 UpdateKids |
| |
|
| parent's SIGUSR1 handler. This handler: |
parent's SIGUSR2 handler. This handler: |
| |
|
| =item |
=item |
| |
|
|
Line 1657 Kills off (via sigint) children for host
|
Line 1582 Kills off (via sigint) children for host
|
| |
|
| =item |
=item |
| |
|
| HUP's children for hosts that already exist (this just forces a status display |
QUITs children for hosts that already exist (this just forces a status display |
| and resets the connection retry count for that host. |
and resets the connection retry count for that host. |
| |
|
| =item |
=item |
|
Line 1668 the start of the master program and main
|
Line 1593 the start of the master program and main
|
| =cut |
=cut |
| |
|
| sub UpdateKids { |
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. |
| |
} |
| |
} |
| } |
} |
| |
|
| |
|
|
Line 1686 sub Restart {
|
Line 1689 sub Restart {
|
| Log("CRITICAL", "Restarting"); |
Log("CRITICAL", "Restarting"); |
| my $execdir = $perlvar{'lonDaemons'}; |
my $execdir = $perlvar{'lonDaemons'}; |
| unlink("$execdir/logs/lonc.pid"); |
unlink("$execdir/logs/lonc.pid"); |
| exec("$execdir/lonc"); |
exec("$execdir/loncnew"); |
| } |
} |
| |
|
| =pod |
=pod |
![[BACK]](/icons/back.gif){kind=icon}
Return to loncnew CVS log ![[TXT]](/icons/text.gif){kind=icon}
![[DIR]](/icons/dir.gif){kind=icon}
Up to [LON-CAPA] / loncom