Update beekeeper scripts for the hive 2.2 -> 2.7 migration

Ensembl · Nov 14, 2024 · bd9e058 · bd9e058
1 parent 333e0fd
commit bd9e058
Show file tree

Hide file tree

Showing 2 changed files with 28 additions and 13 deletions.
diff --git a/tools_hive/utils/beekeeper.pl b/tools_hive/utils/beekeeper.pl
@@ -27,11 +27,6 @@
 $config = eval(uri_unescape($config));
 die "Could not parse command line arguments:\n$@" if $@;
 
-# save pid to pid file as provided
-open(PID, ">$config->{'pid_file'}") or die "Couldn't open $config->{'pid_file'} file: $!";
-print PID $$;
-close PID;
-
 # require SiteDefs and LoadPlugins
 require $config->{'include_script'};
 
@@ -48,10 +43,22 @@
 @INC = grep -d && !$seen{$_} && ($seen{$_} = 1), (reverse(@SiteDefs::ENSEMBL_LIB_DIRS), map("$_/modules", grep /tools/, @{$SiteDefs::ENSEMBL_PLUGINS}), @{$SiteDefs::ENSEMBL_EXTRA_INC}, @INC);
 $ENV{'PERL5LIB'} = join ':', @INC;
 
-# Finally, run the script with correct paths. This is done with 'do', not 'system' so it's run as a part
-# of the same process. We have already saved it's pid in the pid file which can be used to kill the process.
-# It also maintains INCs, but that's not of much use, unfortunately (see above).
-@ARGV = @{$config->{'command_args'}};
-do $config->{'script'};
+# Fork the process
+my $pid = fork();
+
+if (defined $pid) {
+  if ($pid == 0) {
+    # Child process
+    exec($config->{'script'}, @{$config->{'command_args'}}) or die "Couldn't exec: $!";
+  } else {
+    # Parent process
+    # Save the child's PID to pid file using the provided name
+    open(PID, ">$config->{'pid_file'}") or die "Couldn't open $config->{'pid_file'} file: $!";
+    print PID $pid;
+    close PID;
+   }
+} else {
+  die "Fork failed: $!";
+}
 
 # DONE
diff --git a/tools_hive/utils/beekeeper_manager.pl b/tools_hive/utils/beekeeper_manager.pl
@@ -37,16 +37,24 @@
 use Cwd qw(abs_path);
 use URI::Escape qw(uri_escape);
 use Time::localtime;
+use Getopt::Long;
+
+# Configure GetOpt to not panic if it sees options that haven't been explicitly declared
+Getopt::Long::Configure("pass_through");
 
 $Bin                =~ s|\.snapshot?/[^/]+|latest|; #replacing snapshot in path to latest
 my $path            = $Bin;
+
 my $no_cache        = grep { $_ eq '--no_cache_config' }  @ARGV;
 my $redirect_out    = grep { $_ eq '--redirect_output' }  @ARGV;
 my $kill            = grep { $_ eq '--kill'            }  @ARGV;
-my $sleep_time      = grep({ $_ =~ /^\-\-sleep/        }  @ARGV) ? undef : "0.5";
+my $keep_alive      = grep { $_ eq '--keep_alive'      }  @ARGV;
 my $include_script  = abs_path("$Bin/../../../ensembl-webcode/conf/includeSiteDefs.pl");
 my $command_args    = [];
 
+my $sleep_time = 0.5; # default sleep time
+GetOptions("sleep=f" => \$sleep_time);
+
 for (@ARGV) {
   $path = $_ and last if !$path;
   if ($_ =~ /^\-\-path=?(.*)/) {
@@ -117,8 +125,8 @@
 # --sleep arg
 push @$command_args, '--sleep', $sleep_time if $sleep_time;
 
-# any other args in the command line
-push @$command_args, grep($_ !~ /^\-\-(redirect_output|no_cache_config|kill|path)$/, @ARGV);
+# keep beekeeper running
+push @$command_args, '--loop_until', 'FOREVER' if $keep_alive;
 
 # wrapper command
 my $command = sprintf q(perl %s/beekeeper.pl '%s' %s), $Bin, uri_escape(Data::Dumper->new([{