--- loncom/loncron 2005/01/20 16:27:40 1.55 +++ loncom/loncron 2011/11/14 17:27:34 1.95 @@ -2,7 +2,7 @@ # Housekeeping program, started by cron, loncontrol and loncron.pl # -# $Id: loncron,v 1.55 2005/01/20 16:27:40 albertel Exp $ +# $Id: loncron,v 1.95 2011/11/14 17:27:34 raeburn Exp $ # # Copyright Michigan State University Board of Trustees # @@ -32,6 +32,9 @@ use strict; use lib '/home/httpd/lib/perl/'; use LONCAPA::Configuration; +use LONCAPA; +use Apache::lonnet; +use Apache::loncommon; use IO::File; use IO::Socket; @@ -43,21 +46,6 @@ use vars qw (%perlvar %simplestatus $err my $statusdir="/home/httpd/html/lon-status"; -# -------------------------------------------------- Non-critical communication -sub reply { - my ($cmd,$server)=@_; - my $peerfile="$perlvar{'lonSockDir'}/$server"; - my $client=IO::Socket::UNIX->new(Peer =>"$peerfile", - Type => SOCK_STREAM, - Timeout => 10) - or return "con_lost"; - print $client "$cmd\n"; - my $answer=<$client>; - chomp($answer); - if (!$answer) { $answer="con_lost"; } - return $answer; -} - # --------------------------------------------------------- Output error status sub log { @@ -76,30 +64,36 @@ sub errout { ENDERROUT } +sub rotate_logfile { + my ($file,$fh,$description) = @_; + my $size=(stat($file))[7]; + if ($size>40000) { + &log($fh,"
Rotating $description ...
"); + rename("$file.2","$file.3"); + rename("$file.1","$file.2"); + rename("$file","$file.1"); + } +} + sub start_daemon { my ($fh,$daemon,$pidfile,$args) = @_; my $progname=$daemon; - if ($daemon eq 'lonc' && $args eq 'new') { + if ($daemon eq 'lonc') { $progname='loncnew'; - print "new "; } my $error_fname="$perlvar{'lonDaemons'}/logs/${daemon}_errors"; - my $size=(stat($error_fname))[7]; - if ($size>40000) { - &log($fh,"Rotating error logs ...
"); - rename("$error_fname.2","$error_fname.3"); - rename("$error_fname.1","$error_fname.2"); - rename("$error_fname","$error_fname.1"); + &rotate_logfile($error_fname,$fh,'error logs'); + if ($daemon eq 'lonc') { + &clean_sockets($fh); } system("$perlvar{'lonDaemons'}/$progname 2>$perlvar{'lonDaemons'}/logs/${daemon}_errors"); - sleep 2; + sleep 1; if (-e $pidfile) { &log($fh,"Seems like it started ...
"); my $lfh=IO::File->new("$pidfile"); my $daemonpid=<$lfh>; chomp($daemonpid); - sleep 2; - if (kill 0 => $daemonpid) { + if ($daemonpid =~ /^\d+$/ && kill 0 => $daemonpid) { return 1; } else { return 0; @@ -111,10 +105,11 @@ sub start_daemon { } sub checkon_daemon { - my ($fh,$daemon,$maxsize,$sendusr1,$args)=@_; + my ($fh,$daemon,$maxsize,$send,$args)=@_; + my $result; &log($fh,'');
- printf("%-10s ",$daemon);
+ printf("%-15s ",$daemon);
if (-e "$perlvar{'lonDaemons'}/logs/$daemon.log"){
open (DFH,"tail -n25 $perlvar{'lonDaemons'}/logs/$daemon.log|");
while (my $line= Rotating logs ... Cleaned up ".$cleaned." stale sockets. Rotating logs ... Total unsend messages: $unsend$daemon at pid $daemonpid responding");
- if ($sendusr1) { &log($fh,", sending USR1"); }
+ if ($send) { &log($fh,", sending $send"); }
&log($fh,"
");
- if ($sendusr1) { kill USR1 => $daemonpid; }
+ if ($send eq 'USR1') { kill USR1 => $daemonpid; }
+ if ($send eq 'USR2') { kill USR2 => $daemonpid; }
$restartflag=0;
- print "running\n";
+ if ($send eq 'USR2') {
+ $result = 'reloaded';
+ print "reloaded\n";
+ } else {
+ $result = 'running';
+ print "running\n";
+ }
} else {
$errors++;
&log($fh,"$daemon at pid $daemonpid not responding
");
@@ -152,17 +154,20 @@ sub checkon_daemon {
if ($restartflag==1) {
$simplestatus{$daemon}='off';
$errors++;
+ my $kadaemon=$daemon;
+ if ($kadaemon eq 'lonmemcached') { $kadaemon='memcached'; }
&log($fh,'
Killall '.$daemon.': '.
- `killall $daemon 2>&1`.' - ');
- sleep 2;
+ `killall $kadaemon 2>&1`.' - ');
+ sleep 1;
&log($fh,unlink($pidfile).' - '.
- `killall -9 $daemon 2>&1`.
+ `killall -9 $kadaemon 2>&1`.
'
');
&log($fh,"$daemon not running, trying to start
");
if (&start_daemon($fh,$daemon,$pidfile,$args)) {
&log($fh,"$daemon at pid $daemonpid responding
");
$simplestatus{$daemon}='restarted';
+ $result = 'started';
print "started\n";
} else {
$errors++;
@@ -172,8 +177,10 @@ sub checkon_daemon {
if (&start_daemon($fh,$daemon,$pidfile,$args)) {
&log($fh,"$daemon at pid $daemonpid responding
");
$simplestatus{$daemon}='restarted';
+ $result = 'started';
print "started\n";
} else {
+ $result = 'failed';
print " failed\n";
$simplestatus{$daemon}='failed';
$errors++; $errors++;
@@ -196,20 +203,10 @@ sub checkon_daemon {
}
my $fname="$perlvar{'lonDaemons'}/logs/$daemon.log";
-
- my ($dev,$ino,$mode,$nlink,
- $uid,$gid,$rdev,$size,
- $atime,$mtime,$ctime,
- $blksize,$blocks)=stat($fname);
-
- if ($size>$maxsize) {
- &log($fh,"distprobe
");
+ &log($fh,"");
+ open(DSH,"$perlvar{'lonDaemons'}/distprobe |");
+ while (my $line=
");
+
&errout($fh);
}
sub start_logging {
- my ($hostdom,$hostrole,$hostname,$spareid)=@_;
my $fh=IO::File->new(">$statusdir/newstatus.html");
my %simplestatus=();
my $now=time;
@@ -300,7 +306,6 @@ sub start_logging {
Hosts
");
- foreach my $id (sort(keys(%{$hostname}))) {
+ my %hostname = &Apache::lonnet::all_hostnames();
+ foreach my $id (sort(keys(%hostname))) {
+ my $role = (&Apache::lonnet::is_library($id) ? 'library'
+ : 'access');
&log($fh,
- "
\n");
- }
- &log($fh,"$id ".$hostdom->{$id}.
- " ".$hostrole->{$id}.
- " ".$hostname->{$id}." Spare Hosts
");
- foreach my $id (sort(keys(%{$spareid}))) {
- &log($fh,"
\n");
+ &log($fh,"\n");
return $fh;
}
@@ -336,33 +348,127 @@ ENDHEADERS
sub clean_tmp {
my ($fh)=@_;
&log($fh,' \n");
+ }
+ &log($fh,"$id ".&Apache::lonnet::host_domain($id).
+ " ".$role.
+ " ".&Apache::lonnet::hostname($id)." Spare Hosts
");
+ foreach my $type (sort(keys(%Apache::lonnet::spareid))) {
+ &log($fh,"
");
+ foreach my $id (@{ $Apache::lonnet::spareid{$type} }) {
+ &log($fh,"
\nTemporary Files
');
- my $cleaned=0;
- my $old=0;
- while (my $fname=<$perlvar{'lonDaemons'}/tmp/*>) {
- my ($dev,$ino,$mode,$nlink,
- $uid,$gid,$rdev,$size,
- $atime,$mtime,$ctime,
- $blksize,$blocks)=stat($fname);
- my $now=time;
- my $since=$now-$mtime;
- if ($since>$perlvar{'lonExpire'}) {
- my $line='';
- if (open(PROBE,$fname)) {
- $line=
');
+ }
+ }
+ }
+}
+
+sub recursive_clean_tmp {
+ my ($subdir,$cleaned,$old,$removed,$errors) = @_;
+ my $base = "$perlvar{'lonDaemons'}/tmp";
+ my $path = $base;
+ next if ($subdir =~ m{\.\./});
+ next unless (ref($errors) eq 'HASH');
+ unless ($subdir eq '') {
+ $path .= '/'.$subdir;
+ }
+ if (opendir(my $dh,"$path")) {
+ while (my $file = readdir($dh)) {
+ next if ($file =~ /^\.\.?$/);
+ my $fname = "$path/$file";
+ if (-d $fname) {
+ my $innerdir;
+ if ($subdir eq '') {
+ $innerdir = $file;
+ } else {
+ $innerdir = $subdir.'/'.$file;
+ }
+ ($cleaned,$old,$removed) =
+ &recursive_clean_tmp($innerdir,$cleaned,$old,$removed,$errors);
+ my @doms = &Apache::lonnet::current_machine_domains();
+
+ if (open(my $dirhandle,$fname)) {
+ unless (($innerdir eq 'helprequests') ||
+ (($innerdir =~ /^addcourse/) && ($innerdir !~ m{/\d+$}))) {
+ my @contents = grep {!/^\.\.?$/} readdir($dirhandle);
+ join('&&',@contents)."\n";
+ if (scalar(grep {!/^\.\.?$/} readdir($dirhandle)) == 0) {
+ closedir($dirhandle);
+ if ($fname =~ m{^\Q$perlvar{'lonDaemons'}\E/tmp/}) {
+ if (rmdir($fname)) {
+ $removed ++;
+ } elsif (ref($errors->{dir}) eq 'ARRAY') {
+ push(@{$errors->{dir}},$fname);
+ }
+ }
+ }
+ } else {
+ closedir($dirhandle);
+ }
+ }
+ } else {
+ my ($dev,$ino,$mode,$nlink,
+ $uid,$gid,$rdev,$size,
+ $atime,$mtime,$ctime,
+ $blksize,$blocks)=stat($fname);
+ my $now=time;
+ my $since=$now-$mtime;
+ if ($since>$perlvar{'lonExpire'}) {
+ if ($subdir eq '') {
+ my $line='';
+ if ($fname =~ /\.db$/) {
+ if (unlink($fname)) {
+ $cleaned++;
+ } elsif (ref($errors->{file}) eq 'ARRAY') {
+ push(@{$errors->{file}},$fname);
+ }
+ } elsif (open(PROBE,$fname)) {
+ my $line='';
+ $line=$active open session(s)
");
}
+# ----------------------------------------------------------- clean out sockets
+sub clean_sockets {
+ my ($fh)=@_;
+ my $cleaned=0;
+ opendir(SOCKETS,$perlvar{'lonSockDir'});
+ while (my $fname=readdir(SOCKETS)) {
+ next if (-d $fname
+ || $fname=~/(mysqlsock|maximasock|rsock|\Q$perlvar{'lonSockDir'}\E)/);
+ $cleaned++;
+ &log($fh,"Unlinking $fname
");
+ unlink("/home/httpd/sockets/$fname");
+ }
+ &log($fh,"httpd
Access Log
');
-
- open (DFH,"tail -n25 /etc/httpd/logs/access_log|");
- while (my $line=
Error Log
");
-
- open (DFH,"tail -n25 /etc/httpd/logs/error_log|");
- while (my $line=
");
&errout($fh);
}
@@ -436,34 +549,39 @@ sub rotate_lonnet_logs {
} else { &log($fh,"No perm log\n") }
my $fname="$perlvar{'lonDaemons'}/logs/lonnet.log";
-
- my ($dev,$ino,$mode,$nlink,
- $uid,$gid,$rdev,$size,
- $atime,$mtime,$ctime,
- $blksize,$blocks)=stat($fname);
-
- if ($size>40000) {
- &log($fh,"Connections
');
print "testing connections\n";
&log($fh,"");
my ($good,$bad)=(0,0);
- foreach my $tryserver (sort(keys(%{$hostname}))) {
+ my %hostname = &Apache::lonnet::all_hostnames();
+ foreach my $tryserver (sort(keys(%hostname))) {
print(".");
my $result;
- my $answer=reply("ping",$tryserver);
+ my $answer=&Apache::lonnet::reply("ping",$tryserver);
if ($answer eq "$tryserver:$perlvar{'lonHostID'}") {
$result="ok";
$good++;
@@ -509,17 +627,45 @@ sub check_delayed_msg {
}
&log($fh,"
Outgoing Buffer
\n");
-
+# list directory with delayed messages and remember offline servers
+ my %servers=();
open (DFH,"ls -lF $perlvar{'lonSockDir'}/delayed|");
- while (my $line=
\n");
close (DFH);
+ my %hostname = &Apache::lonnet::all_hostnames();
+ my $numhosts = scalar(keys(%hostname));
+# pong to all servers that have delayed messages
+# this will trigger a reverse connection, which should flush the buffers
+ foreach my $tryserver (sort(keys(%servers))) {
+ if ($hostname{$tryserver} || !$numhosts) {
+ my $answer;
+ eval {
+ local $SIG{ ALRM } = sub { die "TIMEOUT" };
+ alarm(20);
+ $answer = &Apache::lonnet::reply("pong",$tryserver);
+ alarm(0);
+ };
+ if ($@ && $@ =~ m/TIMEOUT/) {
+ &log($fh,"Attempted pong to $tryserver timed out
");
+ print "time out while contacting: $tryserver for pong\n";
+ } else {
+ &log($fh,"Pong to $tryserver: $answer
");
+ }
+ } else {
+ &log($fh,"$tryserver has delayed messages, but is not part of the cluster -- skipping 'Pong'.
");
+ }
+ }
}
sub finish_logging {
@@ -541,7 +687,7 @@ sub finish_logging {
}
sub log_simplestatus {
- rename ("$statusdir/newstatus.html","$statusdir/index.html");
+ rename("$statusdir/newstatus.html","$statusdir/index.html");
my $sfh=IO::File->new(">$statusdir/loncron_simple.txt");
foreach (keys %simplestatus) {
@@ -551,15 +697,76 @@ sub log_simplestatus {
$sfh->close();
}
+sub write_loncaparevs {
+ print "Retrieving LON-CAPA version information\n";
+ if (open(my $fh,">$perlvar{'lonTabDir'}/loncaparevs.tab")) {
+ my %hostname = &Apache::lonnet::all_hostnames();
+ foreach my $id (sort(keys(%hostname))) {
+ if ($id ne '') {
+ my $loncaparev;
+ eval {
+ local $SIG{ ALRM } = sub { die "TIMEOUT" };
+ alarm(10);
+ $loncaparev =
+ &Apache::lonnet::get_server_loncaparev('',$id,1,'loncron');
+ alarm(0);
+ };
+ if ($@ && $@ =~ m/TIMEOUT/) {
+ print "time out while contacting lonHost: $id for version\n";
+ }
+ if ($loncaparev =~ /^[\w.\-]+$/) {
+ print $fh $id.':'.$loncaparev."\n";
+ }
+ }
+ }
+ close($fh);
+ }
+ return;
+}
+
+sub write_serverhomeIDs {
+ print "Retrieving LON-CAPA lonHostID information\n";
+ if (open(my $fh,">$perlvar{'lonTabDir'}/serverhomeIDs.tab")) {
+ my %name_to_host = &Apache::lonnet::all_names();
+ foreach my $name (sort(keys(%name_to_host))) {
+ if ($name ne '') {
+ if (ref($name_to_host{$name}) eq 'ARRAY') {
+ my $serverhomeID;
+ eval {
+ local $SIG{ ALRM } = sub { die "TIMEOUT" };
+ alarm(10);
+ $serverhomeID =
+ &Apache::lonnet::get_server_homeID($name,1,'loncron');
+ alarm(0);
+ };
+ if ($@ && $@ =~ m/TIMEOUT/) {
+ print "Time out while contacting server: $name\n";
+ }
+ if ($serverhomeID ne '') {
+ print $fh $name.':'.$serverhomeID."\n";
+ } else {
+ print $fh $name.':'.$name_to_host{$name}->[0]."\n";
+ }
+ }
+ }
+ }
+ close($fh);
+ }
+ return;
+}
+
sub send_mail {
print "sending mail\n";
- my $emailto="$perlvar{'lonAdmEMail'}";
+ my $defdom = $perlvar{'lonDefDomain'};
+ my $origmail = $perlvar{'lonAdmEMail'};
+ my $emailto = &Apache::loncommon::build_recipient_list(undef,
+ 'lonstatusmail',$defdom,$origmail);
if ($totalcount>2500) {
$emailto.=",$perlvar{'lonSysEMail'}";
}
my $subj="LON: $perlvar{'lonHostID'} E:$errors W:$warnings N:$notices";
- my $result=system("metasend -b -t $emailto -s '$subj' -f $statusdir/index.html -m text/html >& /dev/null");
+ my $result=system("metasend -b -S 4000000 -t $emailto -s '$subj' -f $statusdir/index.html -m text/html >& /dev/null");
if ($result != 0) {
$result=system("mail -s '$subj' $emailto < $statusdir/index.html");
}
@@ -570,8 +777,7 @@ sub usage {
loncron - housekeeping program that checks up on various parts of Lon-CAPA
Options:
- --help Display help
- --oldlonc When starting the lonc daemon use 'lonc' not 'loncnew'
+ --help Display
--noemail Do not send the status email
--justcheckconnections Only check the current status of the lonc/d
connections, do not send emails do not
@@ -581,18 +787,23 @@ Options:
running, do not send emails do not
check the lonc/d connections, do not
generate lon-status
+ --justreload Only tell the daemons to reload the config files,
+ do not send emails do not
+ check if the daemons are running, do not
+ generate lon-status
USAGE
}
# ================================================================ Main Program
sub main () {
- my ($oldlonc,$help,$justcheckdaemons,$noemail,$justcheckconnections);
+ my ($help,$justcheckdaemons,$noemail,$justcheckconnections,
+ $justreload);
&GetOptions("help" => \$help,
- "oldlonc" => \$oldlonc,
"justcheckdaemons" => \$justcheckdaemons,
"noemail" => \$noemail,
- "justcheckconnections" => \$justcheckconnections
+ "justcheckconnections" => \$justcheckconnections,
+ "justreload" => \$justreload
);
if ($help) { &usage(); return; }
# --------------------------------- Read loncapa_apache.conf and loncapa.conf
@@ -601,7 +812,7 @@ sub main () {
undef $perlvarref;
delete $perlvar{'lonReceipt'}; # remove since sensitive and not needed
delete $perlvar{'lonSqlAccess'}; # remove since sensitive and not needed
-
+ chdir($perlvar{'lonDaemons'});
# --------------------------------------- Make sure that LON-CAPA is configured
# I only test for one thing here (lonHostID). This is just a safeguard.
if ('{[[[[lonHostID]]]]}' eq $perlvar{'lonHostID'}) {
@@ -627,31 +838,31 @@ sub main () {
exit 1;
}
-# ------------------------------------------------------------- Read hosts file
- my $config=IO::File->new("$perlvar{'lonTabDir'}/hosts.tab");
-
- my (%hostname,%hostdom,%hostrole,%spareid);
- while (my $configline=<$config>) {
- next if ($configline =~ /^(\#|\s*\$)/);
- my ($id,$domain,$role,$name,$ip,$domdescr)=split(/:/,$configline);
- if ($id && $domain && $role && $name && $ip) {
- $hostname{$id}=$name;
- $hostdom{$id}=$domain;
- $hostrole{$id}=$role;
- }
- }
- undef $config;
-
-# ------------------------------------------------------ Read spare server file
- $config=IO::File->new("$perlvar{'lonTabDir'}/spare.tab");
-
- while (my $configline=<$config>) {
- chomp($configline);
- if (($configline) && ($configline ne $perlvar{'lonHostID'})) {
- $spareid{$configline}=1;
- }
+# -------------------------------------------- Force reload of host information
+ &Apache::lonnet::load_hosts_tab(1);
+ &Apache::lonnet::load_domain_tab(1);
+ &Apache::lonnet::get_iphost(1);
+
+# ----------------------------------------- Force firewall update for lond port
+
+ if ((!$justcheckdaemons) && (!$justreload)) {
+ my $now = time;
+ my $tmpfile = $perlvar{'lonDaemons'}.'/tmp/lciptables_iphost_'.
+ $now.$$.int(rand(10000));
+ if (open(my $fh,">$tmpfile")) {
+ my %iphosts = &Apache::lonnet::get_iphost();
+ foreach my $key (keys(%iphosts)) {
+ print $fh "$key\n";
+ }
+ close($fh);
+ if (&LONCAPA::try_to_lock('/tmp/lock_lciptables')) {
+ my $execpath = $perlvar{'lonDaemons'}.'/lciptables';
+ system("$execpath $tmpfile");
+ unlink('/tmp/lock_lciptables'); # Remove the lock file.
+ }
+ unlink($tmpfile);
+ }
}
- undef $config;
# ---------------------------------------------------------------- Start report
@@ -661,30 +872,39 @@ sub main () {
my $fh;
- if (!$justcheckdaemons && !$justcheckconnections) {
- $fh=&start_logging(\%hostdom,\%hostrole,\%hostname,\%spareid);
+ if (!$justcheckdaemons && !$justcheckconnections && !$justreload) {
+ $fh=&start_logging();
&log_machine_info($fh);
&clean_tmp($fh);
&clean_lonIDs($fh);
&check_httpd_logs($fh);
&rotate_lonnet_logs($fh);
+ &rotate_other_logs($fh);
}
- if (!$justcheckconnections) {
+ if (!$justcheckconnections && !$justreload) {
+ &checkon_daemon($fh,'lonmemcached',40000);
&checkon_daemon($fh,'lonsql',200000);
- &checkon_daemon($fh,'lond',40000,1);
- my $args='new';
- if ($oldlonc) { $args = ''; }
- &checkon_daemon($fh,'lonc',40000,1,$args);
- &checkon_daemon($fh,'lonhttpd',40000);
+ if ( &checkon_daemon($fh,'lond',40000,'USR1') eq 'running') {
+ &checkon_daemon($fh,'lond',40000,'USR2');
+ }
+ &checkon_daemon($fh,'lonc',40000,'USR1');
+ &checkon_daemon($fh,'lonmaxima',40000);
+ &checkon_daemon($fh,'lonr',40000);
+ }
+ if ($justreload) {
+ &checkon_daemon($fh,'lond',40000,'USR2');
+ &checkon_daemon($fh,'lonc',40000,'USR2');
}
- if (!$justcheckdaemons) {
- &test_connections($fh,\%hostname);
+ if ($justcheckconnections) {
+ &test_connections($fh);
}
- if (!$justcheckdaemons && !$justcheckconnections) {
+ if (!$justcheckdaemons && !$justcheckconnections && !$justreload) {
&check_delayed_msg($fh);
&finish_logging($fh);
&log_simplestatus();
+ &write_loncaparevs();
+ &write_serverhomeIDs();
if ($totalcount>200 && !$noemail) { &send_mail(); }
}