/usr/lib/oar/oar_meta_sched is in oar-server 2.5.6-2ubuntu1.
This file is owned by root:root, with mode 0o755.
The actual contents of the file can be viewed below.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 | #!/usr/bin/perl
# $Id$
use strict;
use warnings;
use DBI();
use OAR::IO;
use OAR::Modules::Judas qw(oar_debug oar_warn oar_error set_current_log_category);
use OAR::Conf qw(init_conf dump_conf get_conf is_conf get_conf_with_default_param);
use Data::Dumper;
use OAR::Tools;
use OAR::Modules::Hulot;
use OAR::Schedulers::GanttHoleStorage;
# Log category
set_current_log_category('scheduler');
my $exit_code = 0;
my $base = OAR::IO::connect();
my $base_ro = OAR::IO::connect_ro();
init_conf($ENV{OARCONFFILE});
my $Order_part = get_conf("SCHEDULER_RESOURCE_ORDER");
my $binpath;
if (defined($ENV{OARDIR})){
$binpath = $ENV{OARDIR}."/";
}else{
die("[MetaSched] Error: OARDIR env variable must be defined\n");
}
my $Cpuset_field = get_conf("JOB_RESOURCE_MANAGER_PROPERTY_DB_FIELD");
my $Remote_host = get_conf("SERVER_HOSTNAME");
my $Remote_port = get_conf("SERVER_PORT");
my @Sched_available_suspended_resource_type;
my $sched_available_suspended_resource_type_tmp = get_conf("SCHEDULER_AVAILABLE_SUSPENDED_RESOURCE_TYPE");
if (!defined($sched_available_suspended_resource_type_tmp)){
push(@Sched_available_suspended_resource_type, "default");
}else{
@Sched_available_suspended_resource_type = split(" ",$sched_available_suspended_resource_type_tmp);
}
# Look at resources that we must add for each job
my $Resources_to_always_add_type = get_conf("SCHEDULER_RESOURCES_ALWAYS_ASSIGNED_TYPE");
my @Resources_to_always_add = ();
# Do we log every scheduler's computation ?
my $Log_scheduling = get_conf("SCHEDULER_LOG_DECISIONS");
if (defined($Log_scheduling) and $Log_scheduling ne "yes") { $Log_scheduling = undef; }
#minimum of seconds between each jobs
my $Security_time_overhead = 1;
if (is_conf("SCHEDULER_JOB_SECURITY_TIME")){
$Security_time_overhead = get_conf("SCHEDULER_JOB_SECURITY_TIME");
}
my $Openssh_cmd = get_conf("OPENSSH_CMD");
$Openssh_cmd = OAR::Tools::get_default_openssh_cmd() if (!defined($Openssh_cmd));
if (is_conf("OAR_SSH_CONNECTION_TIMEOUT")){
OAR::Tools::set_ssh_timeout(get_conf("OAR_SSH_CONNECTION_TIMEOUT"));
}
my $Server_hostname = get_conf("SERVER_HOSTNAME");
my $Deploy_hostname = get_conf("DEPLOY_HOSTNAME");
if (!defined($Deploy_hostname)){
$Deploy_hostname = $Server_hostname;
}
my $Cosystem_hostname = get_conf("COSYSTEM_HOSTNAME");
if (!defined($Cosystem_hostname)){
$Cosystem_hostname = $Server_hostname;
}
if (is_conf("OAR_RUNTIME_DIRECTORY")){
OAR::Tools::set_default_oarexec_directory(get_conf("OAR_RUNTIME_DIRECTORY"));
}
#minimum of seconds to be considered like a hole in the gantt
my $Minimum_hole_time = 0;
if (is_conf("SCHEDULER_GANTT_HOLE_MINIMUM_TIME")){
$Minimum_hole_time = get_conf("SCHEDULER_GANTT_HOLE_MINIMUM_TIME");
}
# waiting time when a reservation has not all of its nodes
my $Reservation_waiting_timeout = 300;
# global variables : initialized in init_scheduler function
my %besteffort_resource_occupation;
# stock the job ids that where already send to almighty
my %To_launch_jobs_already_treated;
my $current_time_sec = 0;
my $current_time_sql = "0000-00-00 00:00:00";
my $resa_admin_waiting_timeout = get_conf("RESERVATION_WAITING_RESOURCES_TIMEOUT");
oar_debug("[MetaSched] Starting Meta Scheduler\n");
###########################################################################################################
#Initialize Gantt tables with scheduled reservation jobs, Running jobs, toLaunch jobs and Launching jobs; #
###########################################################################################################
# First get resources that we must add for each reservation jobs
if (defined($Resources_to_always_add_type)){
my $tmp_result_state_resources = OAR::IO::get_specific_resource_states($base,$Resources_to_always_add_type);
if (defined($tmp_result_state_resources->{"Alive"})){
@Resources_to_always_add = @{$tmp_result_state_resources->{"Alive"}};
oar_debug("[MetaSched] Got resources to automatically add to every reservations: @Resources_to_always_add\n");
}
}
$Reservation_waiting_timeout = $resa_admin_waiting_timeout if (defined($resa_admin_waiting_timeout));
# Take care of the currently (or nearly) running jobs
# Lock to prevent bipbip update in same time
OAR::IO::lock_table($base,["jobs","assigned_resources","gantt_jobs_predictions","gantt_jobs_resources","job_types","moldable_job_descriptions","resources","job_state_logs","gantt_jobs_predictions_log","gantt_jobs_resources_log"]);
#calculate now date with no overlap with other jobs
my $previous_ref_time_sec = OAR::IO::get_gantt_date($base);
$current_time_sec = OAR::IO::get_date($base);
if ($current_time_sec < $previous_ref_time_sec){
# The system is very fast!!!
my $tmp_delta = $previous_ref_time_sec - $current_time_sec;
if ($tmp_delta > 60){
oar_error("[OAR::Schedulers::Scheduler] init_scheduler : The previous scheduler step was performed $tmp_delta seconds in the future. It is bad, maybe the system clock of the DB server has changed. This can results in bad scheduling decisions.\n");
}else{
$current_time_sec = $previous_ref_time_sec;
if ($tmp_delta > 1){
# Avoid race condition that could increment too much the current time indefinitely
oar_debug("[OAR::Schedulers::Scheduler] init_scheduler : The previous scheduler step was performed $tmp_delta seconds in the future. So we have to wait or the scheduling decisions will be bad\n");
sleep($tmp_delta);
}
}
}
$current_time_sec++;
$current_time_sql = OAR::IO::local_to_sql($current_time_sec);
my ($Alive_resource_id_vec, undef) = OAR::IO::get_resource_ids_in_state($base,"Alive");
my ($Absent_resource_id_vec, undef) = OAR::IO::get_resource_ids_in_state($base,"Absent");
my ($Suspected_resource_id_vec, undef) = OAR::IO::get_resource_ids_in_state($base,"Suspected");
my ($Dead_resource_id_vec, @Dead_resource_ids) = OAR::IO::get_resource_ids_in_state($base,"Dead");
oar_debug("[MetaSched] Retrieve information for already scheduler reservations from database before flush (keep assign resources)\n");
my $reservation_already_there = OAR::IO::get_waiting_reservations_already_scheduled($base);
OAR::IO::gantt_flush_tables($base, $reservation_already_there, $Log_scheduling);
OAR::IO::set_gantt_date($base,$current_time_sec);
#Init the gantt chart with all resources
oar_debug("[MetaSched] Initialize the gantt structure\n");
my ($Max_resources, $All_resource_list_vec, $Only_default_type_resource_vec) = OAR::IO::get_vecs_resources($base);
my $Gantt = {};
$Gantt->{0}->{""}->{""}->{""} = OAR::Schedulers::GanttHoleStorage::new($Max_resources, $Minimum_hole_time);
OAR::Schedulers::GanttHoleStorage::add_new_resources($Gantt->{0}->{""}->{""}->{""}, $All_resource_list_vec);
oar_debug("[MetaSched] Begin processing of already scheduled reservations (accepted with resources assigned)\n");
# Add already scheduled reservations into the gantt
foreach my $i (keys(%{$reservation_already_there})){
my $container_id = 0;
my $inner_id = 0;
my $placeholder_name = "";
my $allowed_name = "";
my $timesharing_user = "";
my $timesharing_name = "";
my $vec = '';
my $types = OAR::IO::get_job_types_hash($base,$i);
my $job_there = OAR::IO::get_job($base,$i);
foreach my $r (@{$reservation_already_there->{$i}->{resources}}){
vec($vec, $r, 1) = 1;
}
if (defined($types->{inner}) and ($types->{inner} =~ /^(\d+)$/)){
$inner_id = $1;
if (defined($Gantt->{$inner_id}->{""}->{""}->{""})){
oar_debug("[MetaSched] [$i] inner job: using container $inner_id\n");
}else{
oar_debug("[MetaSched] [$i] inner job: container $inner_id does not exist but job is running, use container 0.\n");
$inner_id = 0;
}
}
if (defined($types->{container})){ # A container job cannot be placeholder or allowed or timesharing.
oar_debug("[MetaSched] [$i] Job is ($inner_id,,,)\n");
$container_id = $i;
oar_debug("[MetaSched] [$i] container job: create gantt ($container_id,,,)\n");
$Gantt->{$container_id}->{""}->{""}->{""} =
OAR::Schedulers::GanttHoleStorage::new_with_1_hole($Max_resources, $Minimum_hole_time, $reservation_already_there->{$i}->{start_time}, $reservation_already_there->{$i}->{walltime} + $Security_time_overhead , $vec, $All_resource_list_vec);
} else {
($placeholder_name, $allowed_name, $timesharing_user, $timesharing_name) =
OAR::Schedulers::GanttHoleStorage::manage_gantt_for_timesharing_and_placeholder($Gantt, $job_there->{job_user}, $job_there->{job_name}, $types, $inner_id, "[MetaSched] [$i]");
}
#Fill all other gantts
OAR::Schedulers::GanttHoleStorage::fill_gantts($Gantt, $reservation_already_there->{$i}->{start_time}, $reservation_already_there->{$i}->{walltime} + $Security_time_overhead, $vec, $inner_id ,$placeholder_name, $allowed_name, $timesharing_user, $timesharing_name, "[MetaSched] [$i]");
}
oar_debug("[MetaSched] End processing of already scheduled reservations\n");
oar_debug("[MetaSched] Begin processing of current jobs\n");
my @initial_jobs = OAR::IO::get_jobs_in_multiple_states($base, ["Running","toLaunch","Launching","Finishing","Suspended","Resuming"]);
foreach my $job (@initial_jobs){
my $container_id = 0;
my $inner_id = 0;
my $placeholder_name = "";
my $allowed_name = "";
my $timesharing_user = "";
my $timesharing_name = "";
my $i = $job->{job_id};
next if ($job->{assigned_moldable_job} == 0);
my $mold = OAR::IO::get_current_moldable_job($base,$job->{assigned_moldable_job});
# The list of resources on which the job is running
my ($resource_list_vec, @resource_list) = OAR::IO::get_job_current_resources($base, $job->{assigned_moldable_job},undef);
my $date ;
if ($job->{start_time} == 0) {
$date = $current_time_sec;
}elsif ($job->{start_time} + $mold->{moldable_walltime} < $current_time_sec){
$date = $current_time_sec - $mold->{moldable_walltime};
}else{
$date = $job->{start_time};
}
oar_debug("[MetaSched] [$i] add job in database\n");
OAR::IO::add_gantt_scheduled_jobs($base,$job->{assigned_moldable_job},$date,\@resource_list);
my $types = OAR::IO::get_job_types_hash($base,$job->{job_id});
# Ignore besteffort jobs
if (! defined($types->{besteffort})){
my $job_duration = $mold->{moldable_walltime};
if ($job->{state} eq "Suspended"){
# Remove resources of the type specified in SCHEDULER_AVAILABLE_SUSPENDED_RESOURCE_TYPE
($resource_list_vec, @resource_list) = OAR::IO::get_job_current_resources($base, $job->{assigned_moldable_job},\@Sched_available_suspended_resource_type);
}
if ($job->{suspended} eq "YES"){
# This job was suspended so we must recalculate the walltime
$job_duration += OAR::IO::get_job_suspended_sum_duration($base,$job->{job_id},$current_time_sec);
}
if (defined($types->{inner}) and ($types->{inner} =~ /^(\d+)$/)){
$inner_id = $1;
if (defined($Gantt->{$inner_id}->{""}->{""}->{""})){
oar_debug("[MetaSched] [$i] inner job: using container $inner_id\n");
}else{
oar_debug("[MetaSched] [$i] inner job: container $inner_id does not exist but job is running, use container 0.\n");
$inner_id = 0;
}
}
if (defined($types->{container})){ # A container job cannot be placeholder or allowed or timesharing.
oar_debug("[MetaSched] [$i] Job is ($inner_id,,,)\n");
$container_id = $i;
oar_debug("[MetaSched] [$i] container job: create gantt ($container_id,,,)\n");
$Gantt->{$container_id}->{""}->{""}->{""} =
OAR::Schedulers::GanttHoleStorage::new_with_1_hole($Max_resources, $Minimum_hole_time, $date, $job_duration + $Security_time_overhead, $resource_list_vec, $All_resource_list_vec);
} else {
($placeholder_name, $allowed_name, $timesharing_user, $timesharing_name) =
OAR::Schedulers::GanttHoleStorage::manage_gantt_for_timesharing_and_placeholder($Gantt, $job->{job_user}, $job->{job_name}, $types, $inner_id, "[MetaSched] [$i]");
}
#Fill all other gantts
OAR::Schedulers::GanttHoleStorage::fill_gantts($Gantt, $date, $job_duration + $Security_time_overhead, $resource_list_vec, $inner_id ,$placeholder_name, $allowed_name, $timesharing_user, $timesharing_name, "[MetaSched] [$i]");
}else{
#Stock information about besteffort jobs
foreach my $r (@resource_list){
$besteffort_resource_occupation{$r} = $job;
}
}
}
OAR::IO::unlock_table($base);
oar_debug("[MetaSched] End processing of current jobs\n");
# handle accepted advance reservations which do not have assigned resources, mostly because the prediction tables were completely flushed from outside of OAR
oar_debug("[MetaSched] Begin processing of the accepted reservations without assigned resources\n");
my @Rjobs = OAR::IO::get_waiting_reservation_jobs($base);
foreach my $job (@Rjobs){
next if (defined($reservation_already_there->{$job->{job_id}}));
my $container_id = 0;
my $inner_id = 0;
my $placeholder_name = "";
my $allowed_name = "";
my $timesharing_user = "";
my $timesharing_name = "";
my $i = $job->{job_id};
my $types = OAR::IO::get_job_types_hash($base,$i);
my $job_descriptions = OAR::IO::get_resources_data_structure_current_job($base,$job->{job_id});
# For reservation we take the first moldable job
my $moldable = $job_descriptions->[0];
# Get the list of resources where the reservation will be able to be launched = resources in state Alive, Absent or Suspected
# It is assumed here that Absent and Suspected resources could become Alive by the time of the start of the advance reservation job
my $available_resources_vector = $Alive_resource_id_vec | $Absent_resource_id_vec | $Suspected_resource_id_vec;
if (defined($types->{inner}) and ($types->{inner} =~ /^(\d+)$/)){
$inner_id = $1;
if (defined($Gantt->{$inner_id}->{""}->{""}->{""})){
oar_debug("[MetaSched] [$i] inner job: using container $inner_id\n");
}else{
oar_debug("[MetaSched] [$i] inner job: container $inner_id does not exist.\n");
OAR::IO::set_job_message($base,$i,"Container $inner_id does not exist");
OAR::IO::set_job_scheduler_info($base,$i,"Container $inner_id does not exist");
$inner_id=0;
next;
}
}
if (defined($types->{container})){ # A container job cannot be placeholder or allowed or timesharing.
oar_debug("[MetaSched] [$i] Job is ($inner_id,,,) and is a container (ignoring any timesharing or placeholder)\n");
} else {
($placeholder_name, $allowed_name, $timesharing_user, $timesharing_name) =
OAR::Schedulers::GanttHoleStorage::manage_gantt_for_timesharing_and_placeholder($Gantt, $job->{job_user}, $job->{job_name}, $types, $inner_id, "[MetaSched] [$i]");
}
my $free_resources_vec = OAR::Schedulers::GanttHoleStorage::get_free_resources($Gantt->{$inner_id}->{$allowed_name}->{$timesharing_user}->{$timesharing_name},
$job->{start_time},
$moldable->[1] + $Security_time_overhead,
);
$available_resources_vector &= $free_resources_vec;
my $job_properties = "\'1\'";
if ((defined($job->{properties})) and ($job->{properties} ne "")){
$job_properties = $job->{properties};
}
my $resource_id_used_list_vector = '';
my @tree_list;
foreach my $m (@{$moldable->[0]}){
my $tmp_properties = "\'1\'";
if ((defined($m->{property})) and ($m->{property} ne "")){
$tmp_properties = $m->{property};
}
my $tmp_tree;
$tmp_tree = OAR::IO::get_possible_wanted_resources($base_ro,$available_resources_vector,$resource_id_used_list_vector,\@Dead_resource_ids,"$job_properties AND $tmp_properties", $m->{resources}, "".$Order_part);
$tmp_tree = OAR::Schedulers::ResourceTree::delete_tree_nodes_with_not_enough_resources_and_unnecessary_subtrees($tmp_tree, undef);
push(@tree_list, $tmp_tree);
my @leafs = OAR::Schedulers::ResourceTree::get_tree_leafs($tmp_tree);
foreach my $l (@leafs){
vec($resource_id_used_list_vector, OAR::Schedulers::ResourceTree::get_current_resource_value($l), 1) = 1;
}
}
my @resources;
foreach my $t (@tree_list){
foreach my $r (OAR::Schedulers::ResourceTree::get_tree_leafs($t)){
push(@resources, OAR::Schedulers::ResourceTree::get_current_resource_value($r));
}
}
if ($#resources >= 0){
# We can schedule the job
my $vec = '';
foreach my $r (@resources){
vec($vec, $r, 1) = 1;
}
# Create gantt for the new container
if (defined($types->{container})){
$container_id = $i;
oar_debug("[MetaSched] [$i] container job: create gantt ($container_id,,,)\n");
$Gantt->{$container_id}->{""}->{""}->{""} = OAR::Schedulers::GanttHoleStorage::new_with_1_hole($Max_resources, $Minimum_hole_time, $job->{start_time}, $moldable->[1] + $Security_time_overhead, $vec, $All_resource_list_vec);
}
#Fill all other gantts
OAR::Schedulers::GanttHoleStorage::fill_gantts($Gantt, $job->{start_time}, $moldable->[1] + $Security_time_overhead, $vec, $inner_id, $placeholder_name, $allowed_name, $timesharing_user, $timesharing_name, "[MetaSched] [$i]");
# Update database
push(@resources, @Resources_to_always_add);
oar_debug("[MetaSched] [$i] add job in database\n");
oar_warn("[MetaSched] [$i] resources for reservation were lost, got them back\n");
OAR::IO::add_gantt_scheduled_jobs($base,$moldable->[2],$job->{start_time},\@resources);
OAR::IO::add_new_event($base,"GET_RESOURCES_BACK_FOR_RESERVATION",$job->{job_id},"Resources for reservation were lost, got them back");
OAR::IO::set_job_message($base,$job->{job_id},"Renewed resources for reservation");
}else{
oar_warn("[MetaSched] [$i] Resources for reservation are lost, could not get them back\n");
my $message = "No resource for reservation";
if ($job->{message} ne $message) {
# Sounds like it's the first time we notice that the reservation lost its resources, adding event
OAR::IO::add_new_event($base,"UNABLE_TO_GET_RESOURCE_BACK_FOR_RESERVATION",$job->{job_id},"Resources for reservation are lost, could not get them back");
}
OAR::IO::set_job_message($base,$job->{job_id},$message);
}
}
oar_debug("[MetaSched] End processing of the accepted reservations without assigned resources\n");
###########################################################################################################
# End init scheduler #
###########################################################################################################
# launch right reservation jobs
# arg : queue name
# return 1 if there is at least a job to treate, 2 if besteffort jobs must die
sub treate_waiting_reservation_jobs($){
my $queue_name = shift;
oar_debug("[MetaSched] Queue $queue_name: begin processing of waiting reservations\n");
my $return = 0;
my @arrayJobs = OAR::IO::get_waiting_reservation_jobs_specific_queue($base,$queue_name);
# See if there are reserved jobs to launch
foreach my $job (@arrayJobs){
my $job_descriptions = OAR::IO::get_resources_data_structure_current_job($base,$job->{job_id});
my $moldable = $job_descriptions->[0];
# Test if the job is in the past
if ($current_time_sec > $job->{start_time}+$moldable->[1] ){
oar_warn("[MetaSched] [$job->{job_id}] set job state to Error: reservation expired and couldn't be started\n");
OAR::IO::set_job_state($base, $job->{job_id}, "Error");
OAR::IO::set_job_message($base,$job->{job_id},"Reservation expired and couldn't be started.");
$return = 1;
}
my @resa_alive_resources;
my $jobtypes = OAR::IO::get_job_types_hash($base,$job->{job_id});
my $standbystart = ((defined($jobtypes->{deploy}) and $jobtypes->{deploy} eq "standby")
or (defined($jobtypes->{cosystem}) and $jobtypes->{cosystem} eq "standby")
or (defined($jobtypes->{noop}) and $jobtypes->{noop} eq "standby"));
if ($standbystart){
oar_debug("[MetaSched] [$job->{job_id}] reservation is allowed to start with some resources in standby\n");
@resa_alive_resources = OAR::IO::get_gantt_Alive_or_Standby_resources_for_job($base,$moldable->[2], $job->{start_time} + $moldable->[1] + $Security_time_overhead);
} else {
@resa_alive_resources = OAR::IO::get_gantt_Alive_resources_for_job($base,$moldable->[2]);
}
# test if the job is going to be launched and there is no Alive node
if (($#resa_alive_resources < 0) && ($job->{start_time} <= $current_time_sec)){
oar_warn("[MetaSched] [$job->{job_id}] reservation is waiting because no resource is available\n");
OAR::IO::set_gantt_job_startTime($base,$moldable->[2],$current_time_sec + 1);
}elsif($job->{start_time} <= $current_time_sec){
my @resa_resources = OAR::IO::get_gantt_resources_for_job($base,$moldable->[2]);
if ($job->{start_time} + $Reservation_waiting_timeout > $current_time_sec){
if ($#resa_resources > $#resa_alive_resources){
# we have not the same number of nodes than in the query --> wait the specified timeout
oar_warn("[MetaSched] [$job->{job_id}] reservation is waiting because not all resources are available yet (timeout in ".($job->{start_time} + $Reservation_waiting_timeout - $current_time_sec)." seconds)\n");
OAR::IO::set_gantt_job_startTime($base,$moldable->[2],($current_time_sec + 1));
}
}else{
#Check if resources are in Alive state otherwise remove them, the job is going to be launched
foreach my $r (@resa_resources){
my $resource_info = OAR::IO::get_resource_info($base,$r);
if (not ($resource_info->{state} eq "Alive" or ($standbystart and $resource_info->{state} eq "Absent" and
($job->{start_time} + $moldable->[1] + $Security_time_overhead) < $resource_info->{available_upto}))){
oar_warn("[MetaSched] [$job->{job_id}] drop resource $r of the reservation, because resource is $resource_info->{state})\n");
OAR::IO::remove_gantt_resource_job($base, $moldable->[2], $r);
}
}
if ($#resa_resources > $#resa_alive_resources){
OAR::IO::add_new_event($base,"SCHEDULER_REDUCE_NB_NODES_FOR_RESERVATION",$job->{job_id},"[MetaSched] Reduce the number of resources for the job $job->{job_id}.");
my $n = $#resa_alive_resources + 1;
if ($job->{message} =~ s/R\=\d+/R\=$n/g) {
OAR::IO::set_job_message($base,$job->{job_id},$job->{message});
}
}
}
}
}
oar_debug("[MetaSched] Queue $queue_name: end processing of waiting reservations\n");
return($return);
}
# check for jobs with reservation
# arg : queue name, order_part
# return 1 if there is at least a job to treate else 0
sub check_reservation_jobs($$){
my $queue_name = shift;
my $order_part = shift;
oar_debug("[MetaSched] Queue $queue_name: begin processing of new reservations\n");
my $return = 0;
#Init the gantt chart with all resources
my $Gantt = {};
$Gantt->{0}->{""}->{""}->{""} = OAR::Schedulers::GanttHoleStorage::new($Max_resources, $Minimum_hole_time);
OAR::Schedulers::GanttHoleStorage::add_new_resources($Gantt->{0}->{""}->{""}->{""}, $All_resource_list_vec);
# Find jobs to check
my @jobs_to_sched = OAR::IO::get_waiting_toSchedule_reservation_jobs_specific_queue($base,$queue_name);
if ($#jobs_to_sched >= 0){
oar_debug("[MetaSched] Building the gantt structure and filling in already scheduled jobs\n");
# Build gantt diagram of other jobs
# Take care of currently scheduled jobs except besteffort jobs if queue_name is not besteffort
my ($order, %already_scheduled_jobs) = OAR::IO::get_gantt_scheduled_jobs($base);
foreach my $i (keys(%already_scheduled_jobs)){
my $container_id = 0;
my $inner_id = 0;
my $placeholder_name = "";
my $allowed_name = "";
my $timesharing_user = "";
my $timesharing_name = "";
my $types = OAR::IO::get_job_types_hash($base,$i);
if ((! defined($types->{besteffort})) or ($queue_name eq "besteffort")){
my @resource_list = @{$already_scheduled_jobs{$i}->[3]};
my $resource_list_vec = $already_scheduled_jobs{$i}->[10];
my $job_duration = $already_scheduled_jobs{$i}->[1];
if ($already_scheduled_jobs{$i}->[4] eq "Suspended"){
# Remove resources of the type specified in SCHEDULER_AVAILABLE_SUSPENDED_RESOURCE_TYPE
($resource_list_vec, @resource_list) = OAR::IO::get_job_current_resources($base, $already_scheduled_jobs{$i}->[7],\@Sched_available_suspended_resource_type);
next if ($#resource_list < 0);
}
if ($already_scheduled_jobs{$i}->[8] eq "YES"){
# This job was suspended so we must recalculate the walltime
$job_duration += OAR::IO::get_job_suspended_sum_duration($base,$i,$current_time_sec);
}
if (defined($types->{inner}) and ($types->{inner} =~ /^(\d+)$/)){
$inner_id = $1;
if (defined($Gantt->{$inner_id}->{""}->{""}->{""})){
oar_debug("[MetaSched] [$i] inner job: using container $inner_id\n");
}else{
oar_debug("[MetaSched] [$i] inner job: container $inner_id does not exist but job is running, use container 0.\n");
$inner_id = 0;
}
}
if (defined($types->{container})){ # A container job cannot be placeholder or allowed or timesharing.
oar_debug("[MetaSched] [$i] Job is ($inner_id,,,)\n");
$container_id = $i;
oar_debug("[MetaSched] [$i] container job: create gantt ($container_id,,,)\n");
$Gantt->{$container_id}->{""}->{""}->{""} =
OAR::Schedulers::GanttHoleStorage::new_with_1_hole($Max_resources, $Minimum_hole_time, $already_scheduled_jobs{$i}->[0], $job_duration + $Security_time_overhead, $resource_list_vec, $All_resource_list_vec);
} else {
($placeholder_name, $allowed_name, $timesharing_user, $timesharing_name) =
OAR::Schedulers::GanttHoleStorage::manage_gantt_for_timesharing_and_placeholder($Gantt, $already_scheduled_jobs{$i}->[5], $already_scheduled_jobs{$i}->[6], $types, $inner_id, "[MetaSched] [$i]");
}
#Fill all other gantts
OAR::Schedulers::GanttHoleStorage::fill_gantts($Gantt, $already_scheduled_jobs{$i}->[0], $job_duration + $Security_time_overhead, $resource_list_vec, $inner_id ,$placeholder_name, $allowed_name, $timesharing_user, $timesharing_name, "[MetaSched] [$i]");
}
}
oar_debug("[MetaSched] Try and schedule new reservations of queue $queue_name\n");
foreach my $job (@jobs_to_sched){
my $container_id = 0;
my $inner_id = 0;
my $placeholder_name = "";
my $allowed_name = "";
my $timesharing_user = "";
my $timesharing_name = "";
my $i = $job->{job_id};
my $types = OAR::IO::get_job_types_hash($base,$i);
my $job_descriptions = OAR::IO::get_resources_data_structure_current_job($base,$job->{job_id});
# It is a reservation, we take care only of the first moldable job
my $moldable = $job_descriptions->[0];
my $duration = $moldable->[1];
#look if reservation is too old
if ($current_time_sec >= ($job->{start_time} + $duration)){
oar_warn("[MetaSched] [$i] Canceling job: reservation is too old\n");
OAR::IO::set_job_message($base, $job->{job_id}, "Reservation too old");
OAR::IO::set_job_state($base, $job->{job_id}, "toError");
}else{
if ($job->{start_time} < $current_time_sec){
$job->{start_time} = $current_time_sec;
#OAR::IO::set_running_date_arbitrary($base,$job->{job_id},$current_time_sql);
}
# Get the list of resources where the reservation will be able to be launched = resources in state Alive, Absent or Suspected
# It is assumed here that Absent and Suspected resources could become Alive by the time of the start of the advance reservation job
my $available_resources_vector = $Alive_resource_id_vec | $Absent_resource_id_vec | $Suspected_resource_id_vec;
if (defined($types->{inner}) and ($types->{inner} =~ /^(\d+)$/)){
$inner_id = $1;
if (defined($Gantt->{$inner_id}->{""}->{""}->{""})){
oar_debug("[MetaSched] [$i] inner job: using container $inner_id\n");
}else{
oar_debug("[MetaSched] [$i] inner job: container $inner_id does not exist.\n");
OAR::IO::set_job_message($base,$i,"Container $inner_id does not exist");
OAR::IO::set_job_scheduler_info($base,$i,"Container $inner_id does not exist");
$inner_id=0;
next;
}
}
if (defined($types->{container})){ # A container job cannot be placeholder or allowed or timesharing.
oar_debug("[MetaSched] [$i] Job is ($inner_id,,,) and is a container (ignoring any timesharing or placeholder)\n");
} else {
($placeholder_name, $allowed_name, $timesharing_user, $timesharing_name) =
OAR::Schedulers::GanttHoleStorage::manage_gantt_for_timesharing_and_placeholder($Gantt, $job->{job_user}, $job->{job_name}, $types, $inner_id, "[MetaSched] [$i]");
}
my $job_properties = "\'1\'";
if ((defined($job->{properties})) and ($job->{properties} ne "")){
$job_properties = $job->{properties};
}
#my $resource_id_used_list_vector = '';
my @tree_list;
foreach my $m (@{$moldable->[0]}){
my $tmp_properties = "\'1\'";
if ((defined($m->{property})) and ($m->{property} ne "")){
$tmp_properties = $m->{property};
}
my $tmp_tree = OAR::IO::get_possible_wanted_resources($base_ro,$available_resources_vector,undef,\@Dead_resource_ids,"$job_properties AND $tmp_properties", $m->{resources}, $order_part);
$tmp_tree = OAR::Schedulers::ResourceTree::delete_tree_nodes_with_not_enough_resources($tmp_tree);
push(@tree_list, $tmp_tree);
}
oar_debug("[MetaSched] [$i] Find first hole with gantt($inner_id,$allowed_name,$timesharing_user,$timesharing_name)\n");
my @hole = OAR::Schedulers::GanttHoleStorage::find_first_hole($Gantt->{$inner_id}->{$allowed_name}->{$timesharing_user}->{$timesharing_name},$job->{start_time}, $duration + $Security_time_overhead, \@tree_list, 30);
if ($hole[0] == $job->{start_time}){
# The reservation can be scheduled
my @resources;
foreach my $t (@{$hole[1]}){
foreach my $r (OAR::Schedulers::ResourceTree::get_tree_leafs($t)){
push(@resources, OAR::Schedulers::ResourceTree::get_current_resource_value($r));
}
}
# We can schedule the job
oar_warn("[MetaSched] [$i] advance reservation is validated\n");
my $vec = '';
foreach my $r (@resources){
vec($vec, $r, 1) = 1;
}
# Create gantt for the new container
if (defined($types->{container})){
$container_id = $i;
oar_debug("[MetaSched] [$i] container job: create gantt ($container_id,,,)\n");
$Gantt->{$container_id}->{""}->{""}->{""} = OAR::Schedulers::GanttHoleStorage::new_with_1_hole($Max_resources, $Minimum_hole_time, $job->{start_time}, $duration + $Security_time_overhead, $vec, $All_resource_list_vec);
}
#Fill all other gantts
OAR::Schedulers::GanttHoleStorage::fill_gantts($Gantt, $job->{start_time}, $duration + $Security_time_overhead, $vec, $inner_id, $placeholder_name, $allowed_name, $timesharing_user, $timesharing_name, "[MetaSched] [$i]");
# Update database
push(@resources, @Resources_to_always_add);
oar_debug("[MetaSched] [$i] add job in database\n");
OAR::IO::add_gantt_scheduled_jobs($base,$moldable->[2],$job->{start_time},\@resources);
OAR::IO::set_job_state($base, $job->{job_id}, "toAckReservation");
}else{
oar_warn("[MetaSched] [$i] advance reservation cannot be validated, not enough resources\n");
OAR::IO::set_job_state($base, $job->{job_id}, "toError");
if ($hole[0] == OAR::Schedulers::GanttHoleStorage::get_infinity_value()){
OAR::IO::set_job_message($base, $job->{job_id}, "This reservation cannot run");
}else{
OAR::IO::set_job_message($base, $job->{job_id}, "This reservation could run at ".OAR::IO::local_to_sql($hole[0]));
}
}
}
OAR::IO::set_job_resa_state($base, $job->{job_id}, "Scheduled");
$return = 1;
}
}
oar_debug("[MetaSched] Queue $queue_name: end processing of new reservations\n");
return($return);
}
# Detect if there are besteffort jobs to kill
# return 1 if there is at least 1 job to frag otherwise 0
sub check_jobs_to_kill(){
oar_debug("[MetaSched] Begin precessing of besteffort jobs to kill\n");
my $return = 0;
my %nodes_for_jobs_to_launch = OAR::IO::get_gantt_resources_for_jobs_to_launch($base,$current_time_sec);
my %fragged_jobs = ();
foreach my $r (keys(%nodes_for_jobs_to_launch)){
if (defined($besteffort_resource_occupation{$r})) {
if (OAR::IO::is_timesharing_for_2_jobs($base, $nodes_for_jobs_to_launch{$r}, $besteffort_resource_occupation{$r}->{job_id})) {
oar_debug("[MetaSched] Resource $r is needed for job $nodes_for_jobs_to_launch{$r}, but besteffort job $besteffort_resource_occupation{$r}->{job_id} can live, because timesharing compatible\n");
} else {
unless (defined($fragged_jobs{$besteffort_resource_occupation{$r}->{job_id}})) {
my $skip_kill = 0;
# Check if we must checkpoint the besteffort job
if ($besteffort_resource_occupation{$r}->{checkpoint} > 0){
my $checkpoint_first_date;
foreach my $e (OAR::IO::get_job_events($base, $besteffort_resource_occupation{$r}->{job_id})){
if ($e->{type} eq "CHECKPOINT"){
if ((!defined($checkpoint_first_date)) or ($checkpoint_first_date > $e->{date})){
$checkpoint_first_date = $e->{date};
}
}
}
if ((!defined($checkpoint_first_date)) or ($current_time_sec <= ($checkpoint_first_date+$besteffort_resource_occupation{$r}->{checkpoint}))){
# Retrieve node names used by the job
my @hosts = OAR::IO::get_job_current_hostnames($base,$besteffort_resource_occupation{$r}->{job_id});
my $types = OAR::IO::get_job_types_hash($base,$besteffort_resource_occupation{$r}->{job_id});
if ($#hosts < 0 and not defined($types->{cosystem}) and not defined($types->{deploy})) {
oar_debug("[MetaSched] Not sending checkpoint signal to the job $besteffort_resource_occupation{$r}->{job_id} (job with no host)\n");
} else {
my $host_to_connect = $hosts[0];
if (defined($types->{cosystem})) {
$host_to_connect = $Cosystem_hostname;
}elsif (defined($types->{deploy})){
$host_to_connect = $Deploy_hostname;
}
oar_debug("[MetaSched] Sending checkpoint signal to the job $besteffort_resource_occupation{$r}->{job_id} on host $host_to_connect\n");
OAR::IO::add_new_event($base,"CHECKPOINT",$besteffort_resource_occupation{$r}->{job_id},"User oar (metasched) requested a checkpoint on the job $besteffort_resource_occupation{$r}->{job_id} on $host_to_connect");
$skip_kill = 1;
my $str_comment;
my @exit_codes;
# Timeout the ssh command
eval {
$SIG{ALRM} = sub { die "alarm\n" };
alarm(OAR::Tools::get_ssh_timeout());
@exit_codes = OAR::Tools::signal_oarexec($host_to_connect,$besteffort_resource_occupation{$r}->{job_id},"SIGUSR2",1,$base, $Openssh_cmd, '');
alarm(0);
};
if ($@){
if ($@ eq "alarm\n"){
$str_comment = "[MetaSched] Cannot contact $host_to_connect, operation timouted (".OAR::Tools::get_ssh_timeout()." s). So I cannot send checkpoint signal to the job $besteffort_resource_occupation{$r}->{job_id}";
oar_warn("$str_comment\n");
OAR::IO::add_new_event($base,"CHECKPOINT_ERROR",$besteffort_resource_occupation{$r}->{job_id},$str_comment);
}else{
$str_comment = "[MetaSched] An unknown error occured during the sending of the checkpoint signal to the job $besteffort_resource_occupation{$r}->{job_id} on the host $host_to_connect";
oar_warn("$str_comment\n");
OAR::IO::add_new_event($base,"CHECKPOINT_ERROR",$besteffort_resource_occupation{$r}->{job_id},$str_comment);
}
}else{
if ($exit_codes[0] == 0){
$str_comment = "[MetaSched] The job $besteffort_resource_occupation{$r}->{job_id} was notified to checkpoint itself on the node $host_to_connect";
oar_debug("$str_comment\n");
OAR::IO::add_new_event($base,"CHECKPOINT_SUCCESSFULL",$besteffort_resource_occupation{$r}->{job_id},$str_comment);
}else{
$str_comment = "[MetaSched] The kill command return a bad exit code (@exit_codes) for the job $besteffort_resource_occupation{$r}->{job_id} on the node $host_to_connect";
oar_warn("$str_comment\n");
OAR::IO::add_new_event($base,"CHECKPOINT_ERROR",$besteffort_resource_occupation{$r}->{job_id},$str_comment);
}
}
}
}else{
# Retrieve node names used by the job
my @hosts = OAR::IO::get_job_current_hostnames($base,$besteffort_resource_occupation{$r}->{job_id});
my $types = OAR::IO::get_job_types_hash($base,$besteffort_resource_occupation{$r}->{job_id});
if ($#hosts >= 0 or defined($types->{cosystem}) or defined($types->{deploy})) {
my $str_comment = "[MetaSched] Checkpoint timeout for job $besteffort_resource_occupation{$r}->{job_id} ($current_time_sec > ($checkpoint_first_date+$besteffort_resource_occupation{$r}->{checkpoint})";
oar_debug("$str_comment\n");
OAR::IO::add_new_event($base,"CHECKPOINT_TIMEOUT",$besteffort_resource_occupation{$r}->{job_id},$str_comment);
}
}
}
if ($skip_kill == 0){
oar_debug("[MetaSched] Resource $r need to be freed for job $nodes_for_jobs_to_launch{$r}: killing besteffort job $besteffort_resource_occupation{$r}->{job_id}\n");
OAR::IO::add_new_event($base,"BESTEFFORT_KILL",$besteffort_resource_occupation{$r}->{job_id},"[MetaSched] kill the besteffort job $besteffort_resource_occupation{$r}->{job_id}");
OAR::IO::lock_table($base,["frag_jobs","event_logs","jobs"]);
OAR::IO::frag_job($base, $besteffort_resource_occupation{$r}->{job_id});
OAR::IO::unlock_table($base);
}
$fragged_jobs{$besteffort_resource_occupation{$r}->{job_id}} = 1;
$return = 1;
}
}
}
}
oar_debug("[MetaSched] End precessing of besteffort jobs to kill\n");
return($return);
}
# Tell Almighty to run a job
sub notify_to_run_job($$){
my ($base,
$job_id) = @_;
if (!defined($To_launch_jobs_already_treated{$job_id})){
if (OAR::IO::is_job_desktop_computing($base, $job_id)){
oar_debug("[MetaSched] $job_id : Desktop computing job, I don't handle it!\n");
}else{
my $err = OAR::Tools::notify_tcp_socket($Remote_host,$Remote_port,"OARRUNJOB_$job_id");
if (!defined($err)){
$To_launch_jobs_already_treated{$job_id} = 1;
oar_debug("[MetaSched] Notify almighty to launch the job $job_id\n");
}else{
oar_warn("[MetaSched] Not able to notify almighty to launch the job $job_id ($err)\n");
}
}
}
}
# Prepare a job to be run by bipbip
sub prepare_job_to_be_launched($$$$$){
my ($base,
$job_id,
$moldable_job_id,
$job_submission_time,
$resources_array_ref) = @_;
my $running_date = $current_time_sec;
if ($running_date < $job_submission_time){
$running_date = $job_submission_time;
}
OAR::IO::set_running_date_arbitrary($base, $job_id, $running_date);
OAR::IO::set_assigned_moldable_job($base, $job_id, $moldable_job_id);
my $insert_from_file = get_conf_with_default_param("INSERTS_FROM_FILE", "no");
if ($insert_from_file eq 'yes') {
OAR::IO::add_resource_job_pairs_from_file($base, $moldable_job_id, $resources_array_ref);
}else{
OAR::IO::add_resource_job_pairs($base, $moldable_job_id, $resources_array_ref);
}
OAR::IO::set_job_state($base, $job_id, "toLaunch");
notify_to_run_job($base, $job_id);
}
# Detect if there are jobs to launch
# return 1 if there is at least 1 job to launch otherwise 0
sub check_jobs_to_launch(){
oar_debug("[MetaSched] Begin processing of jobs to launch (start time <= $current_time_sql)\n");
my $return_code = 0;
my %jobs_to_launch = OAR::IO::get_gantt_jobs_to_launch($base,$current_time_sec);
foreach my $i (keys(%jobs_to_launch)){
oar_debug("[MetaSched] Set job $i state to toLaunch at $current_time_sql\n");
# We must look at reservations to not go after the initial stop time
my $mold = OAR::IO::get_current_moldable_job($base,$jobs_to_launch{$i}->[0]);
my $job = OAR::IO::get_job($base,$i);
if (($job->{reservation} eq "Scheduled") and ($job->{start_time} < $current_time_sec)){
my $max_time = $mold->{moldable_walltime} - ($current_time_sec - $job->{start_time});
OAR::IO::set_moldable_job_max_time($base,$jobs_to_launch{$i}->[0], $max_time);
OAR::IO::set_gantt_job_startTime($base,$jobs_to_launch{$i}->[0],$current_time_sec);
oar_warn("[MetaSched] Reduce walltime of job $i to $max_time (was $mold->{moldable_walltime})\n");
OAR::IO::add_new_event($base,"REDUCE_RESERVATION_WALLTIME",$i,"Change walltime from $mold->{moldable_walltime} to $max_time");
my $w = OAR::IO::duration_to_sql($max_time);
if ($job->{message} =~ s/W\=\d+\:\d+\:\d+/W\=$w/g){
OAR::IO::set_job_message($base,$i,$job->{message});
}
}
prepare_job_to_be_launched($base, $i, $jobs_to_launch{$i}->[0], $job->{submission_time}, $jobs_to_launch{$i}->[1]);
$return_code = 1;
}
oar_debug("[MetaSched] End processing of jobs to launch\n");
return($return_code);
}
my %initial_time = (
"sec" => $current_time_sec,
"sql" => $current_time_sql
);
my @queues = OAR::IO::get_active_queues($base);
my $name;
my $policy;
foreach my $i (@queues){
$name = $i->[0];
$policy = $i->[1];
my $waiting_jobs = OAR::IO::is_waiting_job_specific_queue_present($base,$name);
if ($waiting_jobs == 1){
oar_debug("[MetaSched] Queue $name: Launching scheduler $policy at time $initial_time{sql}\n");
my ($sched_exit_code,$sched_signal_num,$sched_dumped_core) = (-1, -1, -1);
my $sched_pid = open(READSCHED, "$binpath/schedulers/$policy $name $initial_time{sec} \"$initial_time{sql}\" |") or oar_warn("[MetaSched] Cannot run: \"$binpath/schedulers/$policy $name $initial_time{sec} \"$initial_time{sql}\"\"; $!\n");
if (defined($sched_pid)){
while(my $l = <READSCHED>){
chop($l);
oar_debug("[MetaSched] Read on the scheduler output:$l\n");
if ((get_conf_with_default_param("SCHEDULER_LAUNCHER_OPTIMIZATION", "yes") eq "yes") and
($l =~ m/^SCHEDRUN JOB_ID=(\d+) MOLDABLE_JOB_ID=(\d+) RESOURCES=([\d,]+)$/m)){
my $job_id_to_schedule = $1;
my $moldable_job_id_to_schedule = $2;
my @job_resources_to_schedule = split(',',$3);
my $skip = "";
foreach my $r (@job_resources_to_schedule){
if (defined($besteffort_resource_occupation{$r})){
$skip = "a besteffort job needs to be killed first";
last;
}elsif(vec($Absent_resource_id_vec, $r, 1)){
$skip = "some nodes need to be booted first";
}
}
if ($skip eq ""){
oar_debug("[MetaSched] Early launch the job $job_id_to_schedule\n");
my $job = OAR::IO::get_job($base,$job_id_to_schedule);
prepare_job_to_be_launched($base, $job_id_to_schedule, $moldable_job_id_to_schedule, $job->{submission_time}, \@job_resources_to_schedule);
}else{
oar_debug("[MetaSched] Not able to early launch $job_id_to_schedule: $skip\n");
}
}
}
close(READSCHED);
$sched_exit_code = $? >> 8;
$sched_signal_num = $? & 127;
$sched_dumped_core = $? & 128;
}
if ((!defined($sched_pid)) or ($sched_signal_num != 0) or ($sched_dumped_core != 0)){
oar_error("[MetaSched] Error: execution of $policy failed (pid=".(defined($sched_pid)?"undef":$sched_pid)." signal=$sched_signal_num core_dump=$sched_dumped_core). Disabling queue $name (see `oarnotify')\n");
OAR::IO::stop_a_queue($base,$name);
}
if ($sched_exit_code == 1){
$exit_code = 0;
}elsif($sched_exit_code != 0){
oar_error("[MetaSched] Error: Scheduler $policy returned a bad value: $sched_exit_code, at time $initial_time{sec}. Disabling queue $name (see `oarnotify')\n");
OAR::IO::stop_a_queue($base,$name);
}
treate_waiting_reservation_jobs($name);
check_reservation_jobs($name,$Order_part);
}else{
oar_debug("[MetaSched] Queue $name: No job\n");
}
}
if ($exit_code == 0){
if (check_jobs_to_kill() == 1){
# We must kill besteffort jobs
OAR::Tools::notify_tcp_socket($Remote_host,$Remote_port,"ChState");
$exit_code = 2;
}elsif (check_jobs_to_launch() == 1){
$exit_code = 0;
}
}
#Update visu gantt tables
OAR::IO::update_gantt_visualization($base);
OAR::IO::disconnect($base_ro);
# Manage dynamic node feature
my $timeout_cmd = 10;
my $flagHulot=0;
if (is_conf("SCHEDULER_TIMEOUT")){
$timeout_cmd = get_conf("SCHEDULER_TIMEOUT");
}
if ((is_conf("SCHEDULER_NODE_MANAGER_SLEEP_CMD")
or (get_conf("ENERGY_SAVING_INTERNAL") eq "yes" and is_conf("ENERGY_SAVING_NODE_MANAGER_SLEEP_CMD")))
and is_conf("SCHEDULER_NODE_MANAGER_SLEEP_TIME") and is_conf("SCHEDULER_NODE_MANAGER_IDLE_TIME")){
# Look at nodes that are unused for a duration
my $idle_duration = get_conf("SCHEDULER_NODE_MANAGER_IDLE_TIME");
my $sleep_duration = get_conf("SCHEDULER_NODE_MANAGER_SLEEP_TIME");
my %nodes = OAR::IO::search_idle_nodes($base, $current_time_sec);
my $tmp_time = $current_time_sec - $idle_duration;
my @node_halt;
foreach my $n (keys(%nodes)){
if ($nodes{$n} < $tmp_time){
# Search if the node has enough time to sleep
my $tmp = OAR::IO::get_next_job_date_on_node($base,$n);
if (!defined($tmp) or ($tmp - $sleep_duration > $current_time_sec)){
# Search if node has not been woken up recently
my $wakeup_date = OAR::IO::get_last_wake_up_date_of_node($base,$n);
if (!defined($wakeup_date) or ($wakeup_date < $tmp_time)){
push(@node_halt, $n);
}
}
}
}
if ($#node_halt >= 0){
oar_debug("[MetaSched] Powering off some nodes (energy saving): @node_halt\n");
# Using the built-in energy saving module to shut down nodes
if (get_conf_with_default_param("ENERGY_SAVING_INTERNAL", "no") eq "yes") {
if (OAR::Modules::Hulot::halt_nodes(\@node_halt) ) {
oar_error("[MetaSched] Error: communication problem with the energy saving module (Hulot)\n");
}
$flagHulot=1;
}
# Not using the built-in energy saving module to shut down nodes
else {
my $cmd = get_conf("SCHEDULER_NODE_MANAGER_SLEEP_CMD");
if (! defined(OAR::Tools::fork_and_feed_stdin($cmd, $timeout_cmd, \@node_halt))){
oar_error("[MetaSched] Error: command $cmd timeouted (${timeout_cmd}s) while trying to poweroff some nodes\n");
}
}
}
}
if (is_conf("SCHEDULER_NODE_MANAGER_WAKE_UP_CMD") or (get_conf("ENERGY_SAVING_INTERNAL") eq "yes" and is_conf("ENERGY_SAVING_NODE_MANAGER_WAKE_UP_CMD"))){
# Get nodes which the scheduler wants to schedule jobs to, but which are in the Absent state, to wake them up.
my $wakeup_time = get_conf_with_default_param("SCHEDULER_NODE_MANAGER_WAKEUP_TIME", 1);
my @nodes = OAR::IO::get_gantt_hostname_to_wake_up($base, $current_time_sec, $wakeup_time);
if ($#nodes >= 0){
oar_debug("[MetaSched] Waking-up some nodes: @nodes\n");
# Using the built-in energy saving module to wake up nodes
if (get_conf_with_default_param("ENERGY_SAVING_INTERNAL", "no") eq "yes") {
if (OAR::Modules::Hulot::wake_up_nodes(\@nodes) ) {
oar_error("[MetaSched] Error: Communication problem with the energy saving module (Hulot)\n");
}
$flagHulot=1;
}
# Not using the built-in energy saving module to wake up nodes
else {
my $cmd = get_conf("SCHEDULER_NODE_MANAGER_WAKE_UP_CMD");
if (! defined(OAR::Tools::fork_and_feed_stdin($cmd, $timeout_cmd, \@nodes))){
oar_error("[MetaSched] Error: command $cmd timeouted (${timeout_cmd}s) while trying to wake-up some nodes\n");
}
}
} else {
oar_debug("[MetaSched] No node to wake-up\n");
}
}
# Send CHECK signal to Hulot if needed
if(!$flagHulot and (get_conf_with_default_param("ENERGY_SAVING_INTERNAL", "no") eq "yes")){
if (OAR::Modules::Hulot::check() ) {
oar_error("[MetaSched] Error: communication problem with the energy saving module (Hulot)\n");
}
}
# Search jobs to resume
foreach my $j (OAR::IO::get_jobs_in_state($base, "Resuming")){
my @other_jobs = OAR::IO::get_jobs_on_resuming_job_resources($base,$j->{job_id});
# TODO : look for timesharing other jobs. What do we do?????
if ($#other_jobs < 0){
# We can resume the job
oar_debug("[MetaSched] [$j->{job_id}] Resuming job\n");
my $jobtypes = OAR::IO::get_job_types_hash($base,$j->{job_id});
if (defined($jobtypes->{noop})){
OAR::IO::resume_job_action($base,$j->{job_id});
oar_debug("[MetaSched] [$j->{job_id}] Resume NOOP job OK\n");
}else{
###############
# RESUME PART #
###############
my $script = get_conf("JUST_BEFORE_RESUME_EXEC_FILE");
my $timeout = get_conf("SUSPEND_RESUME_SCRIPT_TIMEOUT");
$timeout = OAR::Tools::get_default_suspend_resume_script_timeout() if (!defined($timeout));
my $skip = 0;
if (defined($script)){
# Launch admin script
my $script_error = 0;
eval {
$SIG{ALRM} = sub { die "alarm\n" };
alarm($timeout);
oar_debug("[MetaSched] [$j->{job_id}] Running post suspend script: `$script $j->{job_id}'\n");
$script_error = system("$script script $j->{job_id}");
oar_debug("[MetaSched] [$j->{job_id}] `$script $j->{job_id}' terminated\n");
alarm(0);
};
if( $@ || ($script_error != 0)){
my $str = "[MetaSched] [$j->{job_id}] Suspend script error: $@; return code = $script_error\n";
oar_error($str);
OAR::IO::add_new_event($base,"RESUME_SCRIPT_ERROR",$j->{job_id},$str);
OAR::IO::frag_job($base,$j->{job_id});
OAR::Tools::notify_tcp_socket($Remote_host,$Remote_port,"Qdel");
$skip = 1;
}
}
if ((defined($Cpuset_field)) and ($skip == 0)){
my $cpuset_name = OAR::IO::get_job_cpuset_name($base, $j->{job_id}) if (defined($Cpuset_field));
my $cpuset_nodes = OAR::IO::get_cpuset_values_for_a_moldable_job($base,$Cpuset_field,$j->{assigned_moldable_job});
my $suspend_data_hash = {
name => $cpuset_name,
job_id => $j->{job_id},
oarexec_pid_file => OAR::Tools::get_oar_pid_file_name($j->{job_id}),
};
if (defined($cpuset_nodes)){
my $taktuk_cmd = get_conf("TAKTUK_CMD");
my $suspend_file = get_conf("SUSPEND_RESUME_FILE");
$suspend_file = OAR::Tools::get_default_suspend_resume_file() if (!defined($suspend_file));
$suspend_file = "$ENV{OARDIR}/$suspend_file" if ($suspend_file !~ /^\//);
my ($tag,@bad) = OAR::Tools::manage_remote_commands([keys(%{$cpuset_nodes})],$suspend_data_hash,$suspend_file,"resume",$Openssh_cmd,$taktuk_cmd,$base);
if ($tag == 0){
my $str = "[MetaSched] [SUSPEND_RESUME] [$j->{job_id}] Bad suspend/resume file: $suspend_file\n";
oar_error($str);
OAR::IO::add_new_event($base, "SUSPEND_RESUME_MANAGER_FILE", $j->{job_id}, $str);
}else{
if (($#bad < 0)){
OAR::IO::resume_job_action($base,$j->{job_id});
}else{
my $str = "[MetaSched] [SUSPEND_RESUME] [$j->{job_id}] Error on several nodes: @bad\n";
oar_error($str);
OAR::IO::add_new_event_with_host($base,"RESUME_ERROR",$j->{job_id},$str,\@bad);
OAR::IO::frag_job($base,$j->{job_id});
# A Leon must be run
$exit_code = 2;
}
}
}
}
#####################
# RESUME PART, END #
#####################
}
}
}
# Notify oarsub -I when they will be launched
foreach my $j (OAR::IO::get_gantt_waiting_interactive_prediction_date($base)){
my ($addr,$port) = split(/:/,$j->{info_type});
my $new_start_prediction = OAR::IO::local_to_sql($j->{start_time});
oar_debug("[MetaSched] [$j->{job_id}] Notifying user of the start prediction: $new_start_prediction (".$j->{message}.")\n");
OAR::Tools::notify_tcp_socket($addr,$port,"[$initial_time{sql}] Start prediction: $new_start_prediction (".$j->{message}.")");
}
# Run the decisions
## Treate "toError" jobs
foreach my $j (OAR::IO::get_jobs_in_state($base,"toError")){
my ($addr,$port) = split(/:/,$j->{info_type});
oar_debug("[MetaSched] Treate job $j->{job_id} in toError state\n");
if (($j->{job_type} eq "INTERACTIVE") or
(($j->{job_type} eq "PASSIVE") and
($j->{reservation} eq "Scheduled"))){
oar_debug("[MetaSched] Notify oarsub job (num:$j->{job_id}) in error; jobInfo=$j->{info_type}\n");
if ((defined(OAR::Tools::notify_tcp_socket($addr,$port,"$j->{message}"))) or
(defined(OAR::Tools::notify_tcp_socket($addr,$port,"BAD JOB")))){
oar_debug("[MetaSched] Cannot open connection to oarsub client for job $j->{job_id} !\n");
}
}
oar_debug("[MetaSched] Set job $j->{job_id} to state Error\n");
OAR::IO::set_job_state($base, $j->{job_id}, "Error");
}
## Treate toAckReservation jobs
foreach my $j (OAR::IO::get_jobs_in_state($base,"toAckReservation")){
my ($addr,$port) = split(/:/,$j->{info_type});
oar_debug("[MetaSched] Treate job $j->{job_id} in toAckReservation state\n");
my $err = OAR::Tools::notify_tcp_socket($addr,$port,"GOOD RESERVATION");
if (defined($err)){
oar_warn("[MetaSched] Frag job $j->{job_id}, I cannot notify oarsub for the reservation ($err)\n");
OAR::IO::add_new_event($base,"CANNOT_NOTIFY_OARSUB",$j->{job_id},"[MetaSched] Can not notify oarsub for the job $j->{job_id} ($err)");
OAR::IO::lock_table($base,["frag_jobs","event_logs","jobs"]);
OAR::IO::frag_job($base,$j->{job_id});
OAR::IO::unlock_table($base);
$exit_code = 2;
}else{
oar_debug("[MetaSched] Notify oarsub for a RESERVATION (idJob=$j->{job_id}) --> OK; jobInfo=$j->{info_type}\n");
OAR::IO::set_job_state($base,$j->{job_id} , "Waiting");
if ($j->{start_time} - 1 <= $current_time_sec){
$exit_code = 1 if ($exit_code == 0);
}
}
}
## Treate toLaunch jobs
foreach my $j (OAR::IO::get_jobs_in_state($base,"toLaunch")){
notify_to_run_job($base, $j->{job_id});
}
OAR::IO::disconnect($base);
oar_debug("[MetaSched] End of Meta Scheduler\n");
exit($exit_code);
|