--- loncom/Attic/lonc 2000/12/05 19:03:55 1.10
+++ loncom/Attic/lonc 2001/11/26 22:20:26 1.18
@@ -15,6 +15,8 @@
# 2/8,7/25 Gerd Kortemeyer
# 12/05 Scott Harrison
# 12/05 Gerd Kortemeyer
+# 01/10/01 Scott Harrison
+# 03/14/01,03/15,06/12,11/26 Gerd Kortemeyer
#
# based on nonforker from Perl Cookbook
# - server who multiplexes without forking
@@ -28,6 +30,9 @@ use Fcntl;
use Tie::RefHash;
use Crypt::IDEA;
+my $status='';
+my $lastlog='';
+
# grabs exception and records it to log before exiting
sub catchexception {
my ($signal)=@_;
@@ -35,25 +40,11 @@ sub catchexception {
$SIG{__DIE__}='DEFAULT';
&logthis("CRITICAL: "
."ABNORMAL EXIT. Child $$ for server $wasserver died through "
- ."$signal with this parameter->[$@]");
+ ."\"$signal\" with this parameter->[$@]");
die($@);
}
-# grabs exception and records it to log before exiting
-# NOTE: we must NOT use the regular (non-overrided) die function in
-# the code because a handler CANNOT be attached to it
-# (despite what some of the documentation says about SIG{__DIE__}.
-sub catchdie {
- my ($message)=@_;
- $SIG{'QUIT'}='DEFAULT';
- $SIG{__DIE__}='DEFAULT';
- &logthis("CRITICAL: "
- ."ABNORMAL EXIT. Child $$ for server $wasserver died through "
- ."\_\_DIE\_\_ with this parameter->[$message]");
- die($message);
-}
-
-$childmaxattempts=10;
+$childmaxattempts=5;
# -------------------------------- Set signal handlers to record abnormal exits
@@ -62,8 +53,7 @@ $SIG{__DIE__}=\&catchexception;
# ------------------------------------ Read httpd access.conf and get variables
-open (CONFIG,"/etc/httpd/conf/access.conf")
- || catchdie "Can't read access.conf";
+open (CONFIG,"/etc/httpd/conf/access.conf") || die "Can't read access.conf";
while ($configline=) {
if ($configline =~ /PerlSetVar/) {
@@ -74,6 +64,16 @@ while ($configline=) {
}
close(CONFIG);
+# ----------------------------- Make sure this process is running from user=www
+my $wwwid=getpwnam('www');
+if ($wwwid!=$<) {
+ $emailto="$perlvar{'lonAdmEMail'},$perlvar{'lonSysEMail'}";
+ $subj="LON: $perlvar{'lonHostID'} User ID mismatch";
+ system("echo 'User ID mismatch. lonc must be run as user www.' |\
+ mailto $emailto -s '$subj' > /dev/null");
+ exit 1;
+}
+
# --------------------------------------------- Check if other instance running
my $pidfile="$perlvar{'lonDaemons'}/logs/lonc.pid";
@@ -82,13 +82,12 @@ if (-e $pidfile) {
my $lfh=IO::File->new("$pidfile");
my $pide=<$lfh>;
chomp($pide);
- if (kill 0 => $pide) { catchdie "already running"; }
+ if (kill 0 => $pide) { die "already running"; }
}
# ------------------------------------------------------------- Read hosts file
-open (CONFIG,"$perlvar{'lonTabDir'}/hosts.tab")
- || catchdie "Can't read host file";
+open (CONFIG,"$perlvar{'lonTabDir'}/hosts.tab") || die "Can't read host file";
while ($configline=) {
my ($id,$domain,$role,$name,$ip)=split(/:/,$configline);
@@ -120,7 +119,13 @@ sub REAPER { # ta
sub HUNTSMAN { # signal handler for SIGINT
local($SIG{CHLD}) = 'IGNORE'; # we're going to kill our children
- kill 'INT' => keys %children;
+ map {
+ $wasserver=$children{$_};
+ &status("Closing $wasserver");
+ &logthis('Closing '.$wasserver.': '.&subreply('exit',$wasserver));
+ &status("Kill PID $_ for $wasserver");
+ kill ('INT',$_);
+ } keys %children;
my $execdir=$perlvar{'lonDaemons'};
unlink("$execdir/logs/lonc.pid");
&logthis("CRITICAL: Shutting down");
@@ -129,12 +134,32 @@ sub HUNTSMAN { # si
sub HUPSMAN { # signal handler for SIGHUP
local($SIG{CHLD}) = 'IGNORE'; # we're going to kill our children
- kill 'INT' => keys %children;
+ map {
+ $wasserver=$children{$_};
+ &status("Closing $wasserver");
+ &logthis('Closing '.$wasserver.': '.&subreply('exit',$wasserver));
+ &status("Kill PID $_ for $wasserver");
+ kill ('INT',$_);
+ } keys %children;
&logthis("CRITICAL: Restarting");
+ unlink("$execdir/logs/lonc.pid");
my $execdir=$perlvar{'lonDaemons'};
exec("$execdir/lonc"); # here we go again
}
+sub checkchildren {
+ &initnewstatus();
+ &logstatus();
+ &logthis('Going to check on the children');
+ map {
+ sleep 1;
+ unless (kill 'USR1' => $_) {
+ &logthis ('Child '.$_.' is dead');
+ &logstatus($$.' is dead');
+ }
+ } sort keys %children;
+}
+
sub USRMAN {
&logthis("USR1: Trying to establish connections again");
foreach $thisserver (keys %hostip) {
@@ -144,6 +169,7 @@ sub USRMAN {
." >$answer<");
}
%childatt=();
+ &checkchildren();
}
# -------------------------------------------------- Non-critical communication
@@ -172,6 +198,7 @@ sub logthis {
my $fh=IO::File->new(">>$execdir/logs/lonc.log");
my $now=time;
my $local=localtime($now);
+ $lastlog=$local.': '.$message;
print $fh "$local ($$): $message\n";
}
@@ -184,14 +211,39 @@ sub logperm {
my $fh=IO::File->new(">>$execdir/logs/lonnet.perm.log");
print $fh "$now:$message:$local\n";
}
+# ------------------------------------------------------------------ Log status
+
+sub logstatus {
+ my $docdir=$perlvar{'lonDocRoot'};
+ my $fh=IO::File->new(">>$docdir/lon-status/loncstatus.txt");
+ print $fh $$."\t".$status."\t".$lastlog."\n";
+}
+
+sub initnewstatus {
+ my $docdir=$perlvar{'lonDocRoot'};
+ my $fh=IO::File->new(">$docdir/lon-status/loncstatus.txt");
+ my $now=time;
+ my $local=localtime($now);
+ print $fh "LONC status $local - parent $$\n\n";
+}
+
+# -------------------------------------------------------------- Status setting
+
+sub status {
+ my $what=shift;
+ my $now=time;
+ my $local=localtime($now);
+ $status=$local.': '.$what;
+}
+
# ---------------------------------------------------- Fork once and dissociate
$fpid=fork;
exit if $fpid;
-catchdie "Couldn't fork: $!" unless defined ($fpid);
+die "Couldn't fork: $!" unless defined ($fpid);
-POSIX::setsid() or catchdie "Can't start new session: $!";
+POSIX::setsid() or die "Can't start new session: $!";
# ------------------------------------------------------- Write our PID on disk
@@ -207,6 +259,8 @@ $SIG{HUP}=$SIG{USR1}='IGNORE';
# Fork off our children, one for every server
+&status("Forking ...");
+
foreach $thisserver (keys %hostip) {
make_new_child($thisserver);
}
@@ -221,11 +275,13 @@ $SIG{USR1} = \&USRMAN;
# And maintain the population.
while (1) {
+ &status("Sleeping");
sleep; # wait for a signal (i.e., child's death)
# See who died and start new one
+ &status("Woke up");
foreach $thisserver (keys %hostip) {
if (!$childpid{$thisserver}) {
- if ($childatt{$thisserver}<=$childmaxattempts) {
+ if ($childatt{$thisserver}<$childmaxattempts) {
$childatt{$thisserver}++;
&logthis(
"INFO: Trying to reconnect for $thisserver "
@@ -246,31 +302,36 @@ sub make_new_child {
# block signal for fork
$sigset = POSIX::SigSet->new(SIGINT);
sigprocmask(SIG_BLOCK, $sigset)
- or catchdie "Can't block SIGINT for fork: $!\n";
+ or die "Can't block SIGINT for fork: $!\n";
- catchdie "fork: $!" unless defined ($pid = fork);
+ die "fork: $!" unless defined ($pid = fork);
if ($pid) {
# Parent records the child's birth and returns.
sigprocmask(SIG_UNBLOCK, $sigset)
- or catchdie "Can't unblock SIGINT for fork: $!\n";
+ or die "Can't unblock SIGINT for fork: $!\n";
$children{$pid} = $conserver;
$childpid{$conserver} = $pid;
return;
} else {
# Child can *not* return from this subroutine.
$SIG{INT} = 'DEFAULT'; # make SIGINT kill us as it did before
-
+ $SIG{USR1}= \&logstatus;
+
# unblock signals
sigprocmask(SIG_UNBLOCK, $sigset)
- or catchdie "Can't unblock SIGINT for fork: $!\n";
+ or die "Can't unblock SIGINT for fork: $!\n";
# ----------------------------- This is the modified main program of non-forker
$port = "$perlvar{'lonSockDir'}/$conserver";
unlink($port);
+
# ---------------------------------------------------- Client to network server
+
+&status("Opening TCP: $conserver");
+
unless (
$remotesock = IO::Socket::INET->new(PeerAddr => $hostip{$conserver},
PeerPort => $perlvar{'londPort'},
@@ -284,19 +345,31 @@ unless (
exit;
};
# --------------------------------------- Send a ping to make other end do USR1
+
+&status("Init dialogue: $conserver");
+
print $remotesock "init\n";
$answer=<$remotesock>;
print $remotesock "$answer";
$answer=<$remotesock>;
chomp($answer);
&logthis("Init reply for $conserver: >$answer<");
+if ($answer ne 'ok') {
+ my $st=120+int(rand(240));
+ &logthis(
+"WARNING: Init failed $conserver ($st secs)");
+ sleep($st);
+ exit;
+}
sleep 5;
+&status("Ponging $conserver");
print $remotesock "pong\n";
$answer=<$remotesock>;
chomp($answer);
&logthis("Pong reply for $conserver: >$answer<");
# ----------------------------------------------------------- Initialize cipher
+&status("Initialize cipher: $conserver");
print $remotesock "ekey\n";
my $buildkey=<$remotesock>;
my $key=$conserver.$perlvar{'lonHostID'};
@@ -307,7 +380,7 @@ $key=$key.$buildkey.$key.$buildkey.$key.
$key=substr($key,0,32);
my $cipherkey=pack("H32",$key);
if ($cipher=new IDEA $cipherkey) {
- &logthis("Secure connection inititalized: $conserver");
+ &logthis("Secure connection initialized: $conserver");
} else {
my $st=120+int(rand(240));
&logthis(
@@ -318,7 +391,7 @@ if ($cipher=new IDEA $cipherkey) {
}
# ----------------------------------------- We're online, send delayed messages
-
+ &status("Checking for delayed messages");
my @allbuffered;
my $path="$perlvar{'lonSockDir'}/delayed";
opendir(DIRHANDLE,$path);
@@ -326,6 +399,7 @@ if ($cipher=new IDEA $cipherkey) {
closedir(DIRHANDLE);
my $dfname;
map {
+ &status("Sending delayed $conserver $_");
$dfname="$path/$_";
&logthis($dfname);
my $wcmd;
@@ -360,6 +434,7 @@ if ($cipher=new IDEA $cipherkey) {
} @allbuffered;
# ------------------------------------------------------- Listen to UNIX socket
+&status("Opening socket $conserver");
unless (
$server = IO::Socket::UNIX->new(Local => $port,
Type => SOCK_STREAM,
@@ -397,11 +472,11 @@ while (1) {
# check for new information on the connections we have
# anything to read or accept?
- foreach $client ($select->can_read(1)) {
+ foreach $client ($select->can_read(0.1)) {
if ($client == $server) {
# accept a new connection
-
+ &status("Accept new connection: $conserver");
$client = $server->accept();
$select->add($client);
nonblock($client);
@@ -416,6 +491,7 @@ while (1) {
delete $outbuffer{$client};
delete $ready{$client};
+ &status("Idle $conserver");
$select->remove($client);
close $client;
next;
@@ -446,16 +522,21 @@ while (1) {
$rv = $client->send($outbuffer{$client}, 0);
unless (defined $rv) {
# Whine, but move on.
- warn "I was told I could write, but I can't.\n";
+ &logthis("I was told I could write, but I can't.\n");
next;
}
+ $errno=$!;
if (($rv == length $outbuffer{$client}) ||
- ($! == POSIX::EWOULDBLOCK)) {
+ ($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.
+
+ &logthis("Dropping data with ".$errno.": ".
+ length($outbuffer{$client}).", $rv");
+
delete $inbuffer{$client};
delete $outbuffer{$client};
delete $ready{$client};
@@ -495,8 +576,11 @@ sub handle {
}
$request="enc:$cmdlength:$encrequest\n";
}
+ &status("Sending $conserver: $request");
print $remotesock "$request";
+ &status("Waiting for reply from $conserver: $request");
$answer=<$remotesock>;
+ &status("Received reply: $request");
if ($answer) {
if ($answer =~ /^enc/) {
my ($cmd,$cmdlength,$encinput)=split(/:/,$answer);
@@ -518,6 +602,7 @@ sub handle {
# ===================================================== Done processing request
}
delete $ready{$client};
+ &status("Completed $conserver: $request");
# -------------------------------------------------------------- End non-forker
}
# ---------------------------------------------------------- End make_new_child
@@ -530,8 +615,8 @@ sub nonblock {
$flags = fcntl($socket, F_GETFL, 0)
- or catchdie "Can't get flags for socket: $!\n";
+ or die "Can't get flags for socket: $!\n";
fcntl($socket, F_SETFL, $flags | O_NONBLOCK)
- or catchdie "Can't make socket nonblocking: $!\n";
+ or die "Can't make socket nonblocking: $!\n";
}