2 # -*- mode: perl; perl-indent-level: 2; indent-tabs-mode: nil; -*-
6 crunch-job: Execute job steps, save snapshots as requested, collate output.
10 Obtain job details from Arvados, run tasks on compute nodes (typically
11 invoked by scheduler on controller):
13 crunch-job --job x-y-z --git-dir /path/to/repo/.git
15 Obtain job details from command line, run tasks on local machine
16 (typically invoked by application or developer on VM):
18 crunch-job --job '{"script_version":"/path/to/working/tree","script":"scriptname",...}'
20 crunch-job --job '{"repository":"https://github.com/curoverse/arvados.git","script_version":"master","script":"scriptname",...}'
28 If the job is already locked, steal the lock and run it anyway.
32 Path to a .git directory (or a git URL) where the commit given in the
33 job's C<script_version> attribute is to be found. If this is I<not>
34 given, the job's C<repository> attribute will be used.
38 Arvados API authorization token to use during the course of the job.
42 Do not clear per-job/task temporary directories during initial job
43 setup. This can speed up development and debugging when running jobs
48 UUID of the job to run, or a JSON-encoded job resource without a
49 UUID. If the latter is given, a new job object will be created.
53 =head1 RUNNING JOBS LOCALLY
55 crunch-job's log messages appear on stderr along with the job tasks'
56 stderr streams. The log is saved in Keep at each checkpoint and when
59 If the job succeeds, the job's output locator is printed on stdout.
61 While the job is running, the following signals are accepted:
65 =item control-C, SIGINT, SIGQUIT
67 Save a checkpoint, terminate any job tasks that are running, and stop.
71 Save a checkpoint and continue.
75 Refresh node allocation (i.e., check whether any nodes have been added
76 or unallocated) and attributes of the Job record that should affect
77 behavior (e.g., cancel job if cancelled_at becomes non-nil).
85 use POSIX ':sys_wait_h';
86 use POSIX qw(strftime);
87 use Fcntl qw(F_GETFL F_SETFL O_NONBLOCK);
91 use Digest::MD5 qw(md5_hex);
97 use File::Path qw( make_path remove_tree );
99 use constant TASK_TEMPFAIL => 111;
100 use constant EX_TEMPFAIL => 75;
101 use constant EX_RETRY_UNLOCKED => 93;
103 $ENV{"TMPDIR"} ||= "/tmp";
104 unless (defined $ENV{"CRUNCH_TMP"}) {
105 $ENV{"CRUNCH_TMP"} = $ENV{"TMPDIR"} . "/crunch-job";
106 if ($ENV{"USER"} ne "crunch" && $< != 0) {
107 # use a tmp dir unique for my uid
108 $ENV{"CRUNCH_TMP"} .= "-$<";
112 # Create the tmp directory if it does not exist
113 if ( ! -d $ENV{"CRUNCH_TMP"} ) {
114 make_path $ENV{"CRUNCH_TMP"} or die "Failed to create temporary working directory: " . $ENV{"CRUNCH_TMP"};
117 $ENV{"JOB_WORK"} = $ENV{"CRUNCH_TMP"} . "/work";
118 $ENV{"CRUNCH_INSTALL"} = "$ENV{CRUNCH_TMP}/opt";
119 $ENV{"CRUNCH_WORK"} = $ENV{"JOB_WORK"}; # deprecated
120 mkdir ($ENV{"JOB_WORK"});
129 my $docker_bin = "docker.io";
130 my $docker_run_args = "";
131 GetOptions('force-unlock' => \$force_unlock,
132 'git-dir=s' => \$git_dir,
133 'job=s' => \$jobspec,
134 'job-api-token=s' => \$job_api_token,
135 'no-clear-tmp' => \$no_clear_tmp,
136 'resume-stash=s' => \$resume_stash,
137 'docker-bin=s' => \$docker_bin,
138 'docker-run-args=s' => \$docker_run_args,
141 if (defined $job_api_token) {
142 $ENV{ARVADOS_API_TOKEN} = $job_api_token;
145 my $have_slurm = exists $ENV{SLURM_JOB_ID} && exists $ENV{SLURM_NODELIST};
150 $main::ENV{CRUNCH_DEBUG} = 1;
154 $main::ENV{CRUNCH_DEBUG} = 0;
157 my $arv = Arvados->new('apiVersion' => 'v1');
166 if ($jobspec =~ /^[-a-z\d]+$/)
168 # $jobspec is an Arvados UUID, not a JSON job specification
169 $Job = api_call("jobs/get", uuid => $jobspec);
174 $local_job = JSON::decode_json($jobspec);
178 # Make sure our workers (our slurm nodes, localhost, or whatever) are
179 # at least able to run basic commands: they aren't down or severely
182 if (($Job || $local_job)->{docker_image_locator}) {
183 $cmd = [$docker_bin, 'ps', '-q'];
185 Log(undef, "Sanity check is `@$cmd`");
186 my ($exited, $stdout, $stderr) = srun_sync(
187 ["srun", "--nodes=\Q$ENV{SLURM_NNODES}\E", "--ntasks-per-node=1"],
189 {label => "sanity check"});
191 Log(undef, "Sanity check failed: ".exit_status_s($exited));
194 Log(undef, "Sanity check OK");
197 my $User = api_call("users/current");
200 if (!$force_unlock) {
201 # Claim this job, and make sure nobody else does
202 eval { api_call("jobs/lock", uuid => $Job->{uuid}); };
204 Log(undef, "Error while locking job, exiting ".EX_TEMPFAIL);
213 map { croak ("No $_ specified") unless $local_job->{$_} }
214 qw(script script_version script_parameters);
217 $local_job->{'is_locked_by_uuid'} = $User->{'uuid'};
218 $local_job->{'started_at'} = gmtime;
219 $local_job->{'state'} = 'Running';
221 $Job = api_call("jobs/create", job => $local_job);
223 $job_id = $Job->{'uuid'};
225 my $keep_logfile = $job_id . '.log.txt';
226 log_writer_start($keep_logfile);
228 $Job->{'runtime_constraints'} ||= {};
229 $Job->{'runtime_constraints'}->{'max_tasks_per_node'} ||= 0;
230 my $max_ncpus = $Job->{'runtime_constraints'}->{'max_tasks_per_node'};
232 my $gem_versions = `gem list --quiet arvados-cli 2>/dev/null`;
234 $gem_versions =~ s/^arvados-cli \(/ with arvados-cli Gem version(s) /;
235 chomp($gem_versions);
236 chop($gem_versions); # Closing parentheses
241 "running from " . ((-e $0) ? realpath($0) : "stdin") . $gem_versions);
243 Log (undef, "check slurm allocation");
246 # Should use $ENV{SLURM_TASKS_PER_NODE} instead of sinfo? (eg. "4(x3),2,4(x2)")
250 my $localcpus = 0 + `grep -cw ^processor /proc/cpuinfo` || 1;
251 push @sinfo, "$localcpus localhost";
253 if (exists $ENV{SLURM_NODELIST})
255 push @sinfo, `sinfo -h --format='%c %N' --nodes=\Q$ENV{SLURM_NODELIST}\E`;
259 my ($ncpus, $slurm_nodelist) = split;
260 $ncpus = $max_ncpus if $max_ncpus && $ncpus > $max_ncpus;
263 while ($slurm_nodelist =~ s/^([^\[,]+?(\[.*?\])?)(,|$)//)
266 if ($nodelist =~ /\[((\d+)(-(\d+))?(,(\d+)(-(\d+))?)*)\]/)
269 foreach (split (",", $ranges))
282 push @nodelist, map {
284 $n =~ s/\[[-,\d]+\]/$_/;
291 push @nodelist, $nodelist;
294 foreach my $nodename (@nodelist)
296 Log (undef, "node $nodename - $ncpus slots");
297 my $node = { name => $nodename,
299 # The number of consecutive times a task has been dispatched
300 # to this node and failed.
302 # The number of consecutive times that SLURM has reported
303 # a node failure since the last successful task.
305 # Don't dispatch work to this node until this time
306 # (in seconds since the epoch) has passed.
308 foreach my $cpu (1..$ncpus)
310 push @slot, { node => $node,
314 push @node, @nodelist;
319 # Ensure that we get one jobstep running on each allocated node before
320 # we start overloading nodes with concurrent steps
322 @slot = sort { $a->{cpu} <=> $b->{cpu} } @slot;
325 $Job->update_attributes(
326 'tasks_summary' => { 'failed' => 0,
331 Log (undef, "start");
332 $SIG{'INT'} = sub { $main::please_freeze = 1; };
333 $SIG{'QUIT'} = sub { $main::please_freeze = 1; };
334 $SIG{'TERM'} = \&croak;
335 $SIG{'TSTP'} = sub { $main::please_freeze = 1; };
336 $SIG{'ALRM'} = sub { $main::please_info = 1; };
337 $SIG{'CONT'} = sub { $main::please_continue = 1; };
338 $SIG{'HUP'} = sub { $main::please_refresh = 1; };
340 $main::please_freeze = 0;
341 $main::please_info = 0;
342 $main::please_continue = 0;
343 $main::please_refresh = 0;
344 my $jobsteps_must_output_keys = 0; # becomes 1 when any task outputs a key
346 grep { $ENV{$1} = $2 if /^(NOCACHE.*?)=(.*)/ } split ("\n", $$Job{knobs});
347 $ENV{"CRUNCH_JOB_UUID"} = $job_id;
348 $ENV{"JOB_UUID"} = $job_id;
351 my @jobstep_todo = ();
352 my @jobstep_done = ();
353 my @jobstep_tomerge = ();
354 my $jobstep_tomerge_level = 0;
355 my $squeue_checked = 0;
356 my $latest_refresh = scalar time;
360 if (defined $Job->{thawedfromkey})
362 thaw ($Job->{thawedfromkey});
366 my $first_task = api_call("job_tasks/create", job_task => {
367 'job_uuid' => $Job->{'uuid'},
372 push @jobstep, { 'level' => 0,
374 'arvados_task' => $first_task,
376 push @jobstep_todo, 0;
382 must_lock_now("$ENV{CRUNCH_TMP}/.lock", "a job is already running here.");
385 my $build_script = handle_readall(\*DATA);
386 my $nodelist = join(",", @node);
387 my $git_tar_count = 0;
389 if (!defined $no_clear_tmp) {
390 # Find FUSE mounts under $CRUNCH_TMP and unmount them. Then clean
391 # up work directories crunch_tmp/work, crunch_tmp/opt,
394 # TODO: When #5036 is done and widely deployed, we can limit mount's
395 # -t option to simply fuse.keep.
396 my ($exited, $stdout, $stderr) = srun_sync(
397 ["srun", "--nodelist=$nodelist", "-D", $ENV{'TMPDIR'}],
398 ['bash', '-ec', '-o', 'pipefail', 'mount -t fuse,fuse.keep | awk "(index(\$3, \"$CRUNCH_TMP\") == 1){print \$3}" | xargs -r -n 1 fusermount -u -z; sleep 1; rm -rf $JOB_WORK $CRUNCH_INSTALL $CRUNCH_TMP/task $CRUNCH_TMP/src* $CRUNCH_TMP/*.cid'],
399 {label => "clean work dirs"});
401 exit(EX_RETRY_UNLOCKED);
405 # If this job requires a Docker image, install that.
406 my ($docker_locator, $docker_stream, $docker_hash, $docker_limitmem, $dockeruserarg);
407 if ($docker_locator = $Job->{docker_image_locator}) {
408 Log (undef, "Install docker image $docker_locator");
409 ($docker_stream, $docker_hash) = find_docker_image($docker_locator);
412 croak("No Docker image hash found from locator $docker_locator");
414 Log (undef, "docker image hash is $docker_hash");
415 $docker_stream =~ s/^\.//;
416 my $docker_install_script = qq{
417 if ! $docker_bin images -q --no-trunc --all | grep -qxF \Q$docker_hash\E; then
418 arv-get \Q$docker_locator$docker_stream/$docker_hash.tar\E | $docker_bin load
422 my ($exited, $stdout, $stderr) = srun_sync(
423 ["srun", "--nodelist=" . join(',', @node)],
424 ["/bin/bash", "-o", "pipefail", "-ec", $docker_install_script],
425 {label => "load docker image"});
428 exit(EX_RETRY_UNLOCKED);
431 # Determine whether this version of Docker supports memory+swap limits.
432 ($exited, $stdout, $stderr) = srun_sync(
433 ["srun", "--nodelist=" . $node[0]],
434 [$docker_bin, 'run', '--help'],
435 {label => "check --memory-swap feature"});
436 $docker_limitmem = ($stdout =~ /--memory-swap/);
438 # Find a non-root Docker user to use.
439 # Tries the default user for the container, then 'crunch', then 'nobody',
440 # testing for whether the actual user id is non-zero. This defends against
441 # mistakes but not malice, but we intend to harden the security in the future
442 # so we don't want anyone getting used to their jobs running as root in their
444 my @tryusers = ("", "crunch", "nobody");
445 foreach my $try_user (@tryusers) {
448 if ($try_user eq "") {
449 $label = "check whether default user is UID 0";
452 $label = "check whether user '$try_user' is UID 0";
453 $try_user_arg = "--user=$try_user";
455 my ($exited, $stdout, $stderr) = srun_sync(
456 ["srun", "--nodelist=" . $node[0]],
458 "$docker_bin run $docker_run_args $try_user_arg $docker_hash id --user"],
461 if ($exited == 0 && $stdout =~ /^\d+$/ && $stdout > 0) {
462 $dockeruserarg = $try_user_arg;
463 if ($try_user eq "") {
464 Log(undef, "Container will run with default user");
466 Log(undef, "Container will run with $dockeruserarg");
472 if (!defined $dockeruserarg) {
473 croak("Could not find a user in container that is not UID 0 (tried default user, @tryusers) or there was a problem running 'id' in the container.");
476 if ($Job->{arvados_sdk_version}) {
477 # The job also specifies an Arvados SDK version. Add the SDKs to the
478 # tar file for the build script to install.
479 Log(undef, sprintf("Packing Arvados SDK version %s for installation",
480 $Job->{arvados_sdk_version}));
481 add_git_archive("git", "--git-dir=$git_dir", "archive",
482 "--prefix=.arvados.sdk/",
483 $Job->{arvados_sdk_version}, "sdk");
487 if (!defined $git_dir && $Job->{'script_version'} =~ m{^/}) {
488 # If script_version looks like an absolute path, *and* the --git-dir
489 # argument was not given -- which implies we were not invoked by
490 # crunch-dispatch -- we will use the given path as a working
491 # directory instead of resolving script_version to a git commit (or
492 # doing anything else with git).
493 $ENV{"CRUNCH_SRC_COMMIT"} = $Job->{'script_version'};
494 $ENV{"CRUNCH_SRC"} = $Job->{'script_version'};
497 # Resolve the given script_version to a git commit sha1. Also, if
498 # the repository is remote, clone it into our local filesystem: this
499 # ensures "git archive" will work, and is necessary to reliably
500 # resolve a symbolic script_version like "master^".
501 $ENV{"CRUNCH_SRC"} = "$ENV{CRUNCH_TMP}/src";
503 Log (undef, "Looking for version ".$Job->{script_version}." from repository ".$Job->{repository});
505 $ENV{"CRUNCH_SRC_COMMIT"} = $Job->{script_version};
507 # If we're running under crunch-dispatch, it will have already
508 # pulled the appropriate source tree into its own repository, and
509 # given us that repo's path as $git_dir.
511 # If we're running a "local" job, we might have to fetch content
512 # from a remote repository.
514 # (Currently crunch-dispatch gives a local path with --git-dir, but
515 # we might as well accept URLs there too in case it changes its
517 my $repo = $git_dir || $Job->{'repository'};
519 # Repository can be remote or local. If remote, we'll need to fetch it
520 # to a local dir before doing `git log` et al.
523 if ($repo =~ m{://|^[^/]*:}) {
524 # $repo is a git url we can clone, like git:// or https:// or
525 # file:/// or [user@]host:repo.git. Note "user/name@host:foo" is
526 # not recognized here because distinguishing that from a local
527 # path is too fragile. If you really need something strange here,
528 # use the ssh:// form.
529 $repo_location = 'remote';
530 } elsif ($repo =~ m{^\.*/}) {
531 # $repo is a local path to a git index. We'll also resolve ../foo
532 # to ../foo/.git if the latter is a directory. To help
533 # disambiguate local paths from named hosted repositories, this
534 # form must be given as ./ or ../ if it's a relative path.
535 if (-d "$repo/.git") {
536 $repo = "$repo/.git";
538 $repo_location = 'local';
540 # $repo is none of the above. It must be the name of a hosted
542 my $arv_repo_list = api_call("repositories/list",
543 'filters' => [['name','=',$repo]]);
544 my @repos_found = @{$arv_repo_list->{'items'}};
545 my $n_found = $arv_repo_list->{'serverResponse'}->{'items_available'};
547 Log(undef, "Repository '$repo' -> "
548 . join(", ", map { $_->{'uuid'} } @repos_found));
551 croak("Error: Found $n_found repositories with name '$repo'.");
553 $repo = $repos_found[0]->{'fetch_url'};
554 $repo_location = 'remote';
556 Log(undef, "Using $repo_location repository '$repo'");
557 $ENV{"CRUNCH_SRC_URL"} = $repo;
559 # Resolve given script_version (we'll call that $treeish here) to a
560 # commit sha1 ($commit).
561 my $treeish = $Job->{'script_version'};
563 if ($repo_location eq 'remote') {
564 # We minimize excess object-fetching by re-using the same bare
565 # repository in CRUNCH_TMP/.git for multiple crunch-jobs -- we
566 # just keep adding remotes to it as needed.
567 my $local_repo = $ENV{'CRUNCH_TMP'}."/.git";
568 my $gitcmd = "git --git-dir=\Q$local_repo\E";
570 # Set up our local repo for caching remote objects, making
572 if (!-d $local_repo) {
573 make_path($local_repo) or croak("Error: could not create $local_repo");
575 # This works (exits 0 and doesn't delete fetched objects) even
576 # if $local_repo is already initialized:
577 `$gitcmd init --bare`;
579 croak("Error: $gitcmd init --bare exited ".exit_status_s($?));
582 # If $treeish looks like a hash (or abbrev hash) we look it up in
583 # our local cache first, since that's cheaper. (We don't want to
584 # do that with tags/branches though -- those change over time, so
585 # they should always be resolved by the remote repo.)
586 if ($treeish =~ /^[0-9a-f]{7,40}$/s) {
587 # Hide stderr because it's normal for this to fail:
588 my $sha1 = `$gitcmd rev-list -n1 ''\Q$treeish\E 2>/dev/null`;
590 # Careful not to resolve a branch named abcdeff to commit 1234567:
591 $sha1 =~ /^$treeish/ &&
592 $sha1 =~ /^([0-9a-f]{40})$/s) {
594 Log(undef, "Commit $commit already present in $local_repo");
598 if (!defined $commit) {
599 # If $treeish isn't just a hash or abbrev hash, or isn't here
600 # yet, we need to fetch the remote to resolve it correctly.
602 # First, remove all local heads. This prevents a name that does
603 # not exist on the remote from resolving to (or colliding with)
604 # a previously fetched branch or tag (possibly from a different
606 remove_tree("$local_repo/refs/heads", {keep_root => 1});
608 Log(undef, "Fetching objects from $repo to $local_repo");
609 `$gitcmd fetch --no-progress --tags ''\Q$repo\E \Q+refs/heads/*:refs/heads/*\E`;
611 croak("Error: `$gitcmd fetch` exited ".exit_status_s($?));
615 # Now that the data is all here, we will use our local repo for
616 # the rest of our git activities.
620 my $gitcmd = "git --git-dir=\Q$repo\E";
621 my $sha1 = `$gitcmd rev-list -n1 ''\Q$treeish\E`;
622 unless ($? == 0 && $sha1 =~ /^([0-9a-f]{40})$/) {
623 croak("`$gitcmd rev-list` exited "
625 .", '$treeish' not found, giving up");
628 Log(undef, "Version $treeish is commit $commit");
630 if ($commit ne $Job->{'script_version'}) {
631 # Record the real commit id in the database, frozentokey, logs,
632 # etc. -- instead of an abbreviation or a branch name which can
633 # become ambiguous or point to a different commit in the future.
634 if (!$Job->update_attributes('script_version' => $commit)) {
635 croak("Error: failed to update job's script_version attribute");
639 $ENV{"CRUNCH_SRC_COMMIT"} = $commit;
640 add_git_archive("$gitcmd archive ''\Q$commit\E");
643 my $git_archive = combined_git_archive();
644 if (!defined $git_archive) {
645 Log(undef, "Skip install phase (no git archive)");
647 Log(undef, "Warning: This probably means workers have no source tree!");
652 my $install_script_tries_left = 3;
653 for (my $attempts = 0; $attempts < 3; $attempts++) {
654 my @srunargs = ("srun",
655 "--nodelist=$nodelist",
656 "-D", $ENV{'TMPDIR'}, "--job-name=$job_id");
657 my @execargs = ("sh", "-c",
658 "mkdir -p $ENV{CRUNCH_INSTALL} && cd $ENV{CRUNCH_TMP} && perl -");
660 $ENV{"CRUNCH_GIT_ARCHIVE_HASH"} = md5_hex($git_archive);
661 my ($stdout, $stderr);
662 ($exited, $stdout, $stderr) = srun_sync(
663 \@srunargs, \@execargs,
664 {label => "run install script on all workers"},
665 $build_script . $git_archive);
667 my $stderr_anything_from_script = 0;
668 for my $line (split(/\n/, $stderr)) {
669 if ($line !~ /^(srun: error: |starting: \[)/) {
670 $stderr_anything_from_script = 1;
674 last if $exited == 0 || $main::please_freeze;
676 # If the install script fails but doesn't print an error message,
677 # the next thing anyone is likely to do is just run it again in
678 # case it was a transient problem like "slurm communication fails
679 # because the network isn't reliable enough". So we'll just do
680 # that ourselves (up to 3 attempts in total). OTOH, if there is an
681 # error message, the problem is more likely to have a real fix and
682 # we should fail the job so the fixing process can start, instead
683 # of doing 2 more attempts.
684 last if $stderr_anything_from_script;
687 foreach my $tar_filename (map { tar_filename_n($_); } (1..$git_tar_count)) {
688 unlink($tar_filename);
696 foreach (qw (script script_version script_parameters runtime_constraints))
700 (ref($Job->{$_}) ? JSON::encode_json($Job->{$_}) : $Job->{$_}));
702 foreach (split (/\n/, $Job->{knobs}))
704 Log (undef, "knob " . $_);
709 $main::success = undef;
715 my $thisround_succeeded = 0;
716 my $thisround_failed = 0;
717 my $thisround_failed_multiple = 0;
718 my $working_slot_count = scalar(@slot);
720 @jobstep_todo = sort { $jobstep[$a]->{level} <=> $jobstep[$b]->{level}
721 or $a <=> $b } @jobstep_todo;
722 my $level = $jobstep[$jobstep_todo[0]]->{level};
724 my $initial_tasks_this_level = 0;
725 foreach my $id (@jobstep_todo) {
726 $initial_tasks_this_level++ if ($jobstep[$id]->{level} == $level);
729 # If the number of tasks scheduled at this level #T is smaller than the number
730 # of slots available #S, only use the first #T slots, or the first slot on
731 # each node, whichever number is greater.
733 # When we dispatch tasks later, we'll allocate whole-node resources like RAM
734 # based on these numbers. Using fewer slots makes more resources available
735 # to each individual task, which should normally be a better strategy when
736 # there are fewer of them running with less parallelism.
738 # Note that this calculation is not redone if the initial tasks at
739 # this level queue more tasks at the same level. This may harm
740 # overall task throughput for that level.
742 if ($initial_tasks_this_level < @node) {
743 @freeslot = (0..$#node);
744 } elsif ($initial_tasks_this_level < @slot) {
745 @freeslot = (0..$initial_tasks_this_level - 1);
747 @freeslot = (0..$#slot);
749 my $round_num_freeslots = scalar(@freeslot);
751 my %round_max_slots = ();
752 for (my $ii = $#freeslot; $ii >= 0; $ii--) {
753 my $this_slot = $slot[$freeslot[$ii]];
754 my $node_name = $this_slot->{node}->{name};
755 $round_max_slots{$node_name} ||= $this_slot->{cpu};
756 last if (scalar(keys(%round_max_slots)) >= @node);
759 Log(undef, "start level $level with $round_num_freeslots slots");
762 my $progress_is_dirty = 1;
763 my $progress_stats_updated = 0;
765 update_progress_stats();
769 for (my $todo_ptr = 0; $todo_ptr <= $#jobstep_todo; $todo_ptr ++)
771 # Don't create new tasks if we already know the job's final result.
772 last if defined($main::success);
774 my $id = $jobstep_todo[$todo_ptr];
775 my $Jobstep = $jobstep[$id];
776 if ($Jobstep->{level} != $level)
781 pipe $reader{$id}, "writer" or croak("pipe() failed: $!");
782 set_nonblocking($reader{$id});
784 my $childslot = $freeslot[0];
785 my $childnode = $slot[$childslot]->{node};
786 my $childslotname = join (".",
787 $slot[$childslot]->{node}->{name},
788 $slot[$childslot]->{cpu});
790 my $childpid = fork();
793 $SIG{'INT'} = 'DEFAULT';
794 $SIG{'QUIT'} = 'DEFAULT';
795 $SIG{'TERM'} = 'DEFAULT';
797 foreach (values (%reader))
801 fcntl ("writer", F_SETFL, 0) or croak ($!); # no close-on-exec
802 open(STDOUT,">&writer");
803 open(STDERR,">&writer");
808 delete $ENV{"GNUPGHOME"};
809 $ENV{"TASK_UUID"} = $Jobstep->{'arvados_task'}->{'uuid'};
810 $ENV{"TASK_QSEQUENCE"} = $id;
811 $ENV{"TASK_SEQUENCE"} = $level;
812 $ENV{"JOB_SCRIPT"} = $Job->{script};
813 while (my ($param, $value) = each %{$Job->{script_parameters}}) {
814 $param =~ tr/a-z/A-Z/;
815 $ENV{"JOB_PARAMETER_$param"} = $value;
817 $ENV{"TASK_SLOT_NODE"} = $slot[$childslot]->{node}->{name};
818 $ENV{"TASK_SLOT_NUMBER"} = $slot[$childslot]->{cpu};
819 $ENV{"TASK_WORK"} = $ENV{"CRUNCH_TMP"}."/task/$childslotname";
820 $ENV{"HOME"} = $ENV{"TASK_WORK"};
821 $ENV{"TASK_TMPDIR"} = $ENV{"TASK_WORK"}; # deprecated
822 $ENV{"CRUNCH_NODE_SLOTS"} = $round_max_slots{$ENV{TASK_SLOT_NODE}};
823 $ENV{"PATH"} = $ENV{"CRUNCH_INSTALL"} . "/bin:" . $ENV{"PATH"};
825 my $keep_mnt = $ENV{"TASK_WORK"}.".keep";
831 "--nodelist=".$childnode->{name},
832 qw(-n1 -c1 -N1 -D), $ENV{'TMPDIR'},
833 "--job-name=$job_id.$id.$$",
836 my $stdbuf = " stdbuf --output=0 --error=0 ";
838 my $arv_file_cache = "";
839 if (defined($Job->{'runtime_constraints'}->{'keep_cache_mb_per_task'})) {
840 $arv_file_cache = "--file-cache=" . ($Job->{'runtime_constraints'}->{'keep_cache_mb_per_task'} * 1024 * 1024);
844 "if [ -e \Q$ENV{TASK_WORK}\E ]; then rm -rf \Q$ENV{TASK_WORK}\E; fi; "
845 ."mkdir -p \Q$ENV{CRUNCH_TMP}\E \Q$ENV{JOB_WORK}\E \Q$ENV{TASK_WORK}\E \Q$keep_mnt\E "
846 ."&& cd \Q$ENV{CRUNCH_TMP}\E "
847 # These environment variables get used explicitly later in
848 # $command. No tool is expected to read these values directly.
849 .q{&& MEM=$(awk '($1 == "MemTotal:"){print $2}' </proc/meminfo) }
850 .q{&& SWAP=$(awk '($1 == "SwapTotal:"){print $2}' </proc/meminfo) }
851 ."&& MEMLIMIT=\$(( (\$MEM * 95) / ($ENV{CRUNCH_NODE_SLOTS} * 100) )) "
852 ."&& let SWAPLIMIT=\$MEMLIMIT+\$SWAP ";
854 $command .= "&& exec arv-mount --read-write --mount-by-pdh=by_pdh --mount-tmp=tmp --crunchstat-interval=10 --allow-other $arv_file_cache \Q$keep_mnt\E --exec ";
855 $ENV{TASK_KEEPMOUNT} = "$keep_mnt/by_pdh";
856 $ENV{TASK_KEEPMOUNT_TMP} = "$keep_mnt/tmp";
860 my $containername = "$Jobstep->{arvados_task}->{uuid}-$Jobstep->{failures}";
861 my $cidfile = "$ENV{CRUNCH_TMP}/$containername.cid";
862 $command .= "crunchstat -cgroup-root=/sys/fs/cgroup -cgroup-parent=docker -cgroup-cid=$cidfile -poll=10000 ";
863 $command .= "$docker_bin run $docker_run_args --name=$containername --attach=stdout --attach=stderr --attach=stdin -i \Q$dockeruserarg\E --cidfile=$cidfile --sig-proxy ";
864 # We only set memory limits if Docker lets us limit both memory and swap.
865 # Memory limits alone have been supported longer, but subprocesses tend
866 # to get SIGKILL if they exceed that without any swap limit set.
867 # See #5642 for additional background.
868 if ($docker_limitmem) {
869 $command .= "--memory=\${MEMLIMIT}k --memory-swap=\${SWAPLIMIT}k ";
872 # The source tree and $destdir directory (which we have
873 # installed on the worker host) are available in the container,
874 # under the same path.
875 $command .= "--volume=\Q$ENV{CRUNCH_SRC}:$ENV{CRUNCH_SRC}:ro\E ";
876 $command .= "--volume=\Q$ENV{CRUNCH_INSTALL}:$ENV{CRUNCH_INSTALL}:ro\E ";
878 # Currently, we make the "by_pdh" directory in arv-mount's mount
879 # point appear at /keep inside the container (instead of using
880 # the same path as the host like we do with CRUNCH_SRC and
881 # CRUNCH_INSTALL). However, crunch scripts and utilities must
882 # not rely on this. They must use $TASK_KEEPMOUNT.
883 $command .= "--volume=\Q$ENV{TASK_KEEPMOUNT}:/keep:ro\E ";
884 $ENV{TASK_KEEPMOUNT} = "/keep";
886 # Ditto TASK_KEEPMOUNT_TMP, as /keep_tmp.
887 $command .= "--volume=\Q$ENV{TASK_KEEPMOUNT_TMP}:/keep_tmp\E ";
888 $ENV{TASK_KEEPMOUNT_TMP} = "/keep_tmp";
890 # TASK_WORK is almost exactly like a docker data volume: it
891 # starts out empty, is writable, and persists until no
892 # containers use it any more. We don't use --volumes-from to
893 # share it with other containers: it is only accessible to this
894 # task, and it goes away when this task stops.
896 # However, a docker data volume is writable only by root unless
897 # the mount point already happens to exist in the container with
898 # different permissions. Therefore, we [1] assume /tmp already
899 # exists in the image and is writable by the crunch user; [2]
900 # avoid putting TASK_WORK inside CRUNCH_TMP (which won't be
901 # writable if they are created by docker while setting up the
902 # other --volumes); and [3] create $TASK_WORK inside the
903 # container using $build_script.
904 $command .= "--volume=/tmp ";
905 $ENV{"TASK_WORK"} = "/tmp/crunch-job-task-work/$childslotname";
906 $ENV{"HOME"} = $ENV{"TASK_WORK"};
907 $ENV{"TASK_TMPDIR"} = $ENV{"TASK_WORK"}; # deprecated
909 # TODO: Share a single JOB_WORK volume across all task
910 # containers on a given worker node, and delete it when the job
911 # ends (and, in case that doesn't work, when the next job
914 # For now, use the same approach as TASK_WORK above.
915 $ENV{"JOB_WORK"} = "/tmp/crunch-job-work";
917 while (my ($env_key, $env_val) = each %ENV)
919 if ($env_key =~ /^(ARVADOS|CRUNCH|JOB|TASK)_/) {
920 $command .= "--env=\Q$env_key=$env_val\E ";
923 $command .= "--env=\QHOME=$ENV{HOME}\E ";
924 $command .= "\Q$docker_hash\E ";
926 if ($Job->{arvados_sdk_version}) {
928 $command .= "perl - \Q$ENV{CRUNCH_SRC}/crunch_scripts/$Job->{script}\E";
930 $command .= "/bin/sh -c \'python -c " .
931 '"from pkg_resources import get_distribution as get; print \"Using Arvados SDK version\", get(\"arvados-python-client\").version"' .
932 ">&2 2>/dev/null; " .
933 "mkdir -p \"$ENV{JOB_WORK}\" \"$ENV{TASK_WORK}\" && " .
934 "if which stdbuf >/dev/null ; then " .
935 " exec $stdbuf \Q$ENV{CRUNCH_SRC}/crunch_scripts/$Job->{script}\E ;" .
937 " exec \Q$ENV{CRUNCH_SRC}/crunch_scripts/$Job->{script}\E ;" .
942 $command .= "crunchstat -cgroup-root=/sys/fs/cgroup -poll=10000 ";
944 $command .= "perl - $ENV{CRUNCH_SRC}/crunch_scripts/" . $Job->{"script"};
947 my @execargs = ('bash', '-c', $command);
948 srun (\@srunargs, \@execargs, undef, $build_script);
949 # exec() failed, we assume nothing happened.
950 die "srun() failed on build script\n";
953 if (!defined $childpid)
964 jobstepname => "$job_id.$id.$childpid",
966 croak ("assert failed: \$slot[$childslot]->{'pid'} exists") if exists $slot[$childslot]->{pid};
967 $slot[$childslot]->{pid} = $childpid;
969 Log ($id, "job_task ".$Jobstep->{'arvados_task'}->{'uuid'});
970 Log ($id, "child $childpid started on $childslotname");
971 $Jobstep->{starttime} = time;
972 $Jobstep->{node} = $childnode->{name};
973 $Jobstep->{slotindex} = $childslot;
974 delete $Jobstep->{stderr};
975 delete $Jobstep->{finishtime};
976 delete $Jobstep->{tempfail};
978 $Jobstep->{'arvados_task'}->{started_at} = strftime "%Y-%m-%dT%H:%M:%SZ", gmtime($Jobstep->{starttime});
979 $Jobstep->{'arvados_task'}->save;
981 splice @jobstep_todo, $todo_ptr, 1;
984 $progress_is_dirty = 1;
988 ($round_num_freeslots > @freeslot && $todo_ptr+1 > $#jobstep_todo))
990 last THISROUND if $main::please_freeze;
991 if ($main::please_info)
993 $main::please_info = 0;
995 create_output_collection();
997 update_progress_stats();
1002 if (!$gotsome || ($latest_refresh + 2 < scalar time))
1004 check_refresh_wanted();
1006 update_progress_stats();
1008 elsif (time - $progress_stats_updated >= 30 || $progress_is_dirty)
1010 update_progress_stats();
1013 select (undef, undef, undef, 0.1);
1015 $working_slot_count = scalar(grep { $_->{node}->{fail_count} == 0 &&
1016 $_->{node}->{hold_count} < 4 } @slot);
1017 if (($thisround_failed_multiple >= 8 && $thisround_succeeded == 0) ||
1018 ($thisround_failed_multiple >= 16 && $thisround_failed_multiple > $thisround_succeeded))
1020 my $message = "Repeated failure rate too high ($thisround_failed_multiple/"
1021 .($thisround_failed+$thisround_succeeded)
1022 .") -- giving up on this round";
1023 Log (undef, $message);
1027 # move slots from freeslot to holdslot (or back to freeslot) if necessary
1028 for (my $i=$#freeslot; $i>=0; $i--) {
1029 if ($slot[$freeslot[$i]]->{node}->{hold_until} > scalar time) {
1030 push @holdslot, (splice @freeslot, $i, 1);
1033 for (my $i=$#holdslot; $i>=0; $i--) {
1034 if ($slot[$holdslot[$i]]->{node}->{hold_until} <= scalar time) {
1035 push @freeslot, (splice @holdslot, $i, 1);
1039 # give up if no nodes are succeeding
1040 if ($working_slot_count < 1) {
1041 Log(undef, "Every node has failed -- giving up");
1048 push @freeslot, splice @holdslot;
1049 map { $slot[$freeslot[$_]]->{node}->{losing_streak} = 0 } (0..$#freeslot);
1052 Log (undef, "wait for last ".(scalar keys %proc)." children to finish");
1055 if ($main::please_continue) {
1056 $main::please_continue = 0;
1059 $main::please_info = 0, freeze(), create_output_collection(), save_meta(1) if $main::please_info;
1061 if (!reapchildren())
1063 check_refresh_wanted();
1065 update_progress_stats();
1066 select (undef, undef, undef, 0.1);
1067 killem (keys %proc) if $main::please_freeze;
1071 update_progress_stats();
1072 freeze_if_want_freeze();
1075 if (!defined $main::success)
1077 if (!@jobstep_todo) {
1079 } elsif ($working_slot_count < 1) {
1080 save_output_collection();
1082 exit(EX_RETRY_UNLOCKED);
1083 } elsif ($thisround_succeeded == 0 &&
1084 ($thisround_failed == 0 || $thisround_failed > 4)) {
1085 my $message = "stop because $thisround_failed tasks failed and none succeeded";
1086 Log (undef, $message);
1091 goto ONELEVEL if !defined $main::success;
1094 release_allocation();
1096 my $collated_output = save_output_collection();
1097 Log (undef, "finish");
1102 if ($collated_output && $main::success) {
1103 $final_state = 'Complete';
1105 $final_state = 'Failed';
1107 $Job->update_attributes('state' => $final_state);
1109 exit (($final_state eq 'Complete') ? 0 : 1);
1113 sub update_progress_stats
1115 $progress_stats_updated = time;
1116 return if !$progress_is_dirty;
1117 my ($todo, $done, $running) = (scalar @jobstep_todo,
1118 scalar @jobstep_done,
1119 scalar keys(%proc));
1120 $Job->{'tasks_summary'} ||= {};
1121 $Job->{'tasks_summary'}->{'todo'} = $todo;
1122 $Job->{'tasks_summary'}->{'done'} = $done;
1123 $Job->{'tasks_summary'}->{'running'} = $running;
1124 $Job->update_attributes('tasks_summary' => $Job->{'tasks_summary'});
1125 Log (undef, "status: $done done, $running running, $todo todo");
1126 $progress_is_dirty = 0;
1133 my $pid = waitpid (-1, WNOHANG);
1134 return 0 if $pid <= 0;
1136 my $whatslot = ($slot[$proc{$pid}->{slot}]->{node}->{name}
1138 . $slot[$proc{$pid}->{slot}]->{cpu});
1139 my $jobstepidx = $proc{$pid}->{jobstepidx};
1140 my $elapsed = time - $proc{$pid}->{time};
1141 my $Jobstep = $jobstep[$jobstepidx];
1143 my $childstatus = $?;
1144 my $exitvalue = $childstatus >> 8;
1145 my $exitinfo = "exit ".exit_status_s($childstatus);
1146 $Jobstep->{'arvados_task'}->reload;
1147 my $task_success = $Jobstep->{'arvados_task'}->{success};
1149 Log ($jobstepidx, "child $pid on $whatslot $exitinfo success=$task_success");
1151 if (!defined $task_success) {
1152 # task did not indicate one way or the other --> fail
1153 Log($jobstepidx, sprintf(
1154 "ERROR: Task process exited %s, but never updated its task record to indicate success and record its output.",
1155 exit_status_s($childstatus)));
1156 $Jobstep->{'arvados_task'}->{success} = 0;
1157 $Jobstep->{'arvados_task'}->save;
1164 $temporary_fail ||= $Jobstep->{tempfail};
1165 $temporary_fail ||= ($exitvalue == TASK_TEMPFAIL);
1167 ++$thisround_failed;
1168 ++$thisround_failed_multiple if $Jobstep->{'failures'} >= 1;
1170 # Check for signs of a failed or misconfigured node
1171 if (++$slot[$proc{$pid}->{slot}]->{node}->{losing_streak} >=
1172 2+$slot[$proc{$pid}->{slot}]->{node}->{ncpus}) {
1173 # Don't count this against jobstep failure thresholds if this
1174 # node is already suspected faulty and srun exited quickly
1175 if ($slot[$proc{$pid}->{slot}]->{node}->{hold_until} &&
1177 Log ($jobstepidx, "blaming failure on suspect node " .
1178 $slot[$proc{$pid}->{slot}]->{node}->{name});
1179 $temporary_fail ||= 1;
1181 ban_node_by_slot($proc{$pid}->{slot});
1184 Log ($jobstepidx, sprintf('failure (#%d, %s) after %d seconds',
1185 ++$Jobstep->{'failures'},
1186 $temporary_fail ? 'temporary' : 'permanent',
1189 if (!$temporary_fail || $Jobstep->{'failures'} >= 3) {
1190 # Give up on this task, and the whole job
1193 # Put this task back on the todo queue
1194 push @jobstep_todo, $jobstepidx;
1195 $Job->{'tasks_summary'}->{'failed'}++;
1199 ++$thisround_succeeded;
1200 $slot[$proc{$pid}->{slot}]->{node}->{losing_streak} = 0;
1201 $slot[$proc{$pid}->{slot}]->{node}->{hold_until} = 0;
1202 $slot[$proc{$pid}->{slot}]->{node}->{fail_count} = 0;
1203 push @jobstep_done, $jobstepidx;
1204 Log ($jobstepidx, "success in $elapsed seconds");
1206 $Jobstep->{exitcode} = $childstatus;
1207 $Jobstep->{finishtime} = time;
1208 $Jobstep->{'arvados_task'}->{finished_at} = strftime "%Y-%m-%dT%H:%M:%SZ", gmtime($Jobstep->{finishtime});
1209 $Jobstep->{'arvados_task'}->save;
1210 process_stderr_final ($jobstepidx);
1211 Log ($jobstepidx, sprintf("task output (%d bytes): %s",
1212 length($Jobstep->{'arvados_task'}->{output}),
1213 $Jobstep->{'arvados_task'}->{output}));
1215 close $reader{$jobstepidx};
1216 delete $reader{$jobstepidx};
1217 delete $slot[$proc{$pid}->{slot}]->{pid};
1218 push @freeslot, $proc{$pid}->{slot};
1221 if ($task_success) {
1223 my $newtask_list = [];
1224 my $newtask_results;
1226 $newtask_results = api_call(
1229 'created_by_job_task_uuid' => $Jobstep->{'arvados_task'}->{uuid}
1231 'order' => 'qsequence',
1232 'offset' => scalar(@$newtask_list),
1234 push(@$newtask_list, @{$newtask_results->{items}});
1235 } while (@{$newtask_results->{items}});
1236 foreach my $arvados_task (@$newtask_list) {
1238 'level' => $arvados_task->{'sequence'},
1240 'arvados_task' => $arvados_task
1242 push @jobstep, $jobstep;
1243 push @jobstep_todo, $#jobstep;
1247 $progress_is_dirty = 1;
1251 sub check_refresh_wanted
1253 my @stat = stat $ENV{"CRUNCH_REFRESH_TRIGGER"};
1255 $stat[9] > $latest_refresh &&
1256 # ...and we have actually locked the job record...
1257 $job_id eq $Job->{'uuid'}) {
1258 $latest_refresh = scalar time;
1259 my $Job2 = api_call("jobs/get", uuid => $jobspec);
1260 for my $attr ('cancelled_at',
1261 'cancelled_by_user_uuid',
1262 'cancelled_by_client_uuid',
1264 $Job->{$attr} = $Job2->{$attr};
1266 if ($Job->{'state'} ne "Running") {
1267 if ($Job->{'state'} eq "Cancelled") {
1268 Log (undef, "Job cancelled at " . $Job->{'cancelled_at'} . " by user " . $Job->{'cancelled_by_user_uuid'});
1270 Log (undef, "Job state unexpectedly changed to " . $Job->{'state'});
1273 $main::please_freeze = 1;
1280 my $last_squeue_check = $squeue_checked;
1282 # Do not call `squeue` or check the kill list more than once every
1284 return if $last_squeue_check > time - 15;
1285 $squeue_checked = time;
1287 # Look for children from which we haven't received stderr data since
1288 # the last squeue check. If no such children exist, all procs are
1289 # alive and there's no need to even look at squeue.
1291 # As long as the crunchstat poll interval (10s) is shorter than the
1292 # squeue check interval (15s) this should make the squeue check an
1294 my $silent_procs = 0;
1295 for my $procinfo (values %proc)
1297 my $jobstep = $jobstep[$procinfo->{jobstepidx}];
1298 if ($jobstep->{stderr_at} < $last_squeue_check)
1303 return if $silent_procs == 0;
1305 # use killem() on procs whose killtime is reached
1306 while (my ($pid, $procinfo) = each %proc)
1308 my $jobstep = $jobstep[$procinfo->{jobstepidx}];
1309 if (exists $procinfo->{killtime}
1310 && $procinfo->{killtime} <= time
1311 && $jobstep->{stderr_at} < $last_squeue_check)
1314 if ($jobstep->{stderr_at}) {
1315 $sincewhen = " in last " . (time - $jobstep->{stderr_at}) . "s";
1317 Log($procinfo->{jobstepidx}, "killing orphaned srun process $pid (task not in slurm queue, no stderr received$sincewhen)");
1324 # here is an opportunity to check for mysterious problems with local procs
1328 # Get a list of steps still running. Note: squeue(1) says --steps
1329 # selects a format (which we override anyway) and allows us to
1330 # specify which steps we're interested in (which we don't).
1331 # Importantly, it also changes the meaning of %j from "job name" to
1332 # "step name" and (although this isn't mentioned explicitly in the
1333 # docs) switches from "one line per job" mode to "one line per step"
1334 # mode. Without it, we'd just get a list of one job, instead of a
1336 my @squeue = `squeue --jobs=\Q$ENV{SLURM_JOB_ID}\E --steps --format='%j' --noheader`;
1339 Log(undef, "warning: squeue exit status $? ($!)");
1344 # which of my jobsteps are running, according to squeue?
1346 for my $jobstepname (@squeue)
1348 $ok{$jobstepname} = 1;
1351 # Check for child procs >60s old and not mentioned by squeue.
1352 while (my ($pid, $procinfo) = each %proc)
1354 if ($procinfo->{time} < time - 60
1355 && $procinfo->{jobstepname}
1356 && !exists $ok{$procinfo->{jobstepname}}
1357 && !exists $procinfo->{killtime})
1359 # According to slurm, this task has ended (successfully or not)
1360 # -- but our srun child hasn't exited. First we must wait (30
1361 # seconds) in case this is just a race between communication
1362 # channels. Then, if our srun child process still hasn't
1363 # terminated, we'll conclude some slurm communication
1364 # error/delay has caused the task to die without notifying srun,
1365 # and we'll kill srun ourselves.
1366 $procinfo->{killtime} = time + 30;
1367 Log($procinfo->{jobstepidx}, "notice: task is not in slurm queue but srun process $pid has not exited");
1373 sub release_allocation
1377 Log (undef, "release job allocation");
1378 system "scancel $ENV{SLURM_JOB_ID}";
1386 foreach my $jobstepidx (keys %reader)
1389 if ($jobstep[$jobstepidx]->{stdout_r} &&
1390 0 < sysread ($jobstep[$jobstepidx]->{stdout_r}, $buf, 65536))
1392 print STDERR $buf if $ENV{CRUNCH_DEBUG};
1393 if (exists $jobstep[$jobstepidx]->{stdout_captured}) {
1394 $jobstep[$jobstepidx]->{stdout_captured} .= $buf;
1398 if (0 < sysread ($reader{$jobstepidx}, $buf, 65536))
1400 print STDERR $buf if $ENV{CRUNCH_DEBUG};
1401 $jobstep[$jobstepidx]->{stderr_at} = time;
1402 $jobstep[$jobstepidx]->{stderr} .= $buf;
1403 if (exists $jobstep[$jobstepidx]->{stderr_captured}) {
1404 $jobstep[$jobstepidx]->{stderr_captured} .= $buf;
1408 # Consume everything up to the last \n
1409 preprocess_stderr ($jobstepidx);
1411 if (length ($jobstep[$jobstepidx]->{stderr}) > 16384)
1413 # If we get a lot of stderr without a newline, chop off the
1414 # front to avoid letting our buffer grow indefinitely.
1415 substr ($jobstep[$jobstepidx]->{stderr},
1416 0, length($jobstep[$jobstepidx]->{stderr}) - 8192) = "";
1424 # Consume all full lines of stderr for a jobstep. Everything after the
1425 # last newline will remain in $jobstep[$jobstepidx]->{stderr} after
1427 sub preprocess_stderr
1429 my $jobstepidx = shift;
1431 while ($jobstep[$jobstepidx]->{stderr} =~ /^(.*?)\n/) {
1433 substr $jobstep[$jobstepidx]->{stderr}, 0, 1+length($line), "";
1434 Log ($jobstepidx, "stderr $line");
1435 if ($line =~ /srun: error: (SLURM job $ENV{SLURM_JOB_ID} has expired|Unable to confirm allocation for job $ENV{SLURM_JOB_ID})/) {
1437 $main::please_freeze = 1;
1439 elsif (!exists $jobstep[$jobstepidx]->{slotindex}) {
1440 # Skip the following tempfail checks if this srun proc isn't
1441 # attached to a particular worker slot.
1443 elsif ($line =~ /srun: error: (Node failure on|Aborting, .*\bio error\b)/) {
1444 my $job_slot_index = $jobstep[$jobstepidx]->{slotindex};
1445 my $job_slot_index = $jobstep[$jobstepidx]->{slotindex};
1446 $slot[$job_slot_index]->{node}->{fail_count}++;
1447 $jobstep[$jobstepidx]->{tempfail} = 1;
1448 ban_node_by_slot($job_slot_index);
1450 elsif ($line =~ /srun: error: (Unable to create job step|.*: Communication connection failure)/) {
1451 $jobstep[$jobstepidx]->{tempfail} = 1;
1452 ban_node_by_slot($jobstep[$jobstepidx]->{slotindex});
1454 elsif ($line =~ /arvados\.errors\.Keep/) {
1455 $jobstep[$jobstepidx]->{tempfail} = 1;
1461 sub process_stderr_final
1463 my $jobstepidx = shift;
1464 preprocess_stderr ($jobstepidx);
1467 Log ($jobstepidx, "stderr $_");
1468 } split ("\n", $jobstep[$jobstepidx]->{stderr});
1469 $jobstep[$jobstepidx]->{stderr} = '';
1476 if (!open($keep, "-|", "arv-get", "--retries", retry_count(), $hash)) {
1477 Log(undef, "fetch_block run error from arv-get $hash: $!");
1480 my $output_block = "";
1483 my $bytes = sysread($keep, $buf, 1024 * 1024);
1484 if (!defined $bytes) {
1485 Log(undef, "fetch_block read error from arv-get: $!");
1486 $output_block = undef;
1488 } elsif ($bytes == 0) {
1489 # sysread returns 0 at the end of the pipe.
1492 # some bytes were read into buf.
1493 $output_block .= $buf;
1498 Log(undef, "fetch_block arv-get exited " . exit_status_s($?));
1499 $output_block = undef;
1501 return $output_block;
1504 # Create a collection by concatenating the output of all tasks (each
1505 # task's output is either a manifest fragment, a locator for a
1506 # manifest fragment stored in Keep, or nothing at all). Return the
1507 # portable_data_hash of the new collection.
1508 sub create_output_collection
1510 Log (undef, "collate");
1512 my ($child_out, $child_in);
1513 my $pid = open2($child_out, $child_in, 'python', '-c', q{
1516 print (arvados.api("v1").collections().
1517 create(body={"manifest_text": sys.stdin.read()}).
1518 execute(num_retries=int(sys.argv[1]))["portable_data_hash"])
1522 my $manifest_size = 0;
1526 my $output = $_->{'arvados_task'}->{output};
1527 next if (!defined($output));
1529 if ($output =~ /^[0-9a-f]{32}(\+\S+)*$/) {
1530 $next_write = fetch_block($output);
1532 $next_write = $output;
1534 if (defined($next_write)) {
1535 if (!defined(syswrite($child_in, $next_write))) {
1536 # There's been an error writing. Stop the loop.
1537 # We'll log details about the exit code later.
1540 $manifest_size += length($next_write);
1543 my $uuid = $_->{'arvados_task'}->{'uuid'};
1544 Log (undef, "Error retrieving '$output' output by task $task_idx ($uuid)");
1549 Log(undef, "collated output manifest text to send to API server is $manifest_size bytes with access tokens");
1552 my $s = IO::Select->new($child_out);
1553 if ($s->can_read(120)) {
1554 sysread($child_out, $joboutput, 1024 * 1024);
1557 Log(undef, "output collection creation exited " . exit_status_s($?));
1563 Log (undef, "timed out while creating output collection");
1564 foreach my $signal (2, 2, 2, 15, 15, 9) {
1565 kill($signal, $pid);
1566 last if waitpid($pid, WNOHANG) == -1;
1575 # Calls create_output_collection, logs the result, and returns it.
1576 # If that was successful, save that as the output in the job record.
1577 sub save_output_collection {
1578 my $collated_output = create_output_collection();
1580 if (!$collated_output) {
1581 Log(undef, "Failed to write output collection");
1584 Log(undef, "job output $collated_output");
1585 $Job->update_attributes('output' => $collated_output);
1587 return $collated_output;
1594 my $sig = 2; # SIGINT first
1595 if (exists $proc{$_}->{"sent_$sig"} &&
1596 time - $proc{$_}->{"sent_$sig"} > 4)
1598 $sig = 15; # SIGTERM if SIGINT doesn't work
1600 if (exists $proc{$_}->{"sent_$sig"} &&
1601 time - $proc{$_}->{"sent_$sig"} > 4)
1603 $sig = 9; # SIGKILL if SIGTERM doesn't work
1605 if (!exists $proc{$_}->{"sent_$sig"})
1607 Log ($proc{$_}->{jobstepidx}, "sending 2x signal $sig to pid $_");
1609 select (undef, undef, undef, 0.1);
1612 kill $sig, $_; # srun wants two SIGINT to really interrupt
1614 $proc{$_}->{"sent_$sig"} = time;
1615 $proc{$_}->{"killedafter"} = time - $proc{$_}->{"time"};
1625 vec($bits,fileno($_),1) = 1;
1631 # Send log output to Keep via arv-put.
1633 # $log_pipe_in and $log_pipe_out are the input and output filehandles to the arv-put pipe.
1634 # $log_pipe_out_buf is a string containing all output read from arv-put so far.
1635 # $log_pipe_out_select is an IO::Select object around $log_pipe_out.
1636 # $log_pipe_pid is the pid of the arv-put subprocess.
1638 # The only functions that should access these variables directly are:
1640 # log_writer_start($logfilename)
1641 # Starts an arv-put pipe, reading data on stdin and writing it to
1642 # a $logfilename file in an output collection.
1644 # log_writer_read_output([$timeout])
1645 # Read output from $log_pipe_out and append it to $log_pipe_out_buf.
1646 # Passes $timeout to the select() call, with a default of 0.01.
1647 # Returns the result of the last read() call on $log_pipe_out, or
1648 # -1 if read() wasn't called because select() timed out.
1649 # Only other log_writer_* functions should need to call this.
1651 # log_writer_send($txt)
1652 # Writes $txt to the output log collection.
1654 # log_writer_finish()
1655 # Closes the arv-put pipe and returns the output that it produces.
1657 # log_writer_is_active()
1658 # Returns a true value if there is currently a live arv-put
1659 # process, false otherwise.
1661 my ($log_pipe_in, $log_pipe_out, $log_pipe_out_buf, $log_pipe_out_select,
1664 sub log_writer_start($)
1666 my $logfilename = shift;
1667 $log_pipe_pid = open2($log_pipe_out, $log_pipe_in,
1671 '--filename', $logfilename,
1673 $log_pipe_out_buf = "";
1674 $log_pipe_out_select = IO::Select->new($log_pipe_out);
1677 sub log_writer_read_output {
1678 my $timeout = shift || 0.01;
1680 while ($read && $log_pipe_out_select->can_read($timeout)) {
1681 $read = read($log_pipe_out, $log_pipe_out_buf, 65536,
1682 length($log_pipe_out_buf));
1684 if (!defined($read)) {
1685 Log(undef, "error reading log manifest from arv-put: $!");
1690 sub log_writer_send($)
1693 print $log_pipe_in $txt;
1694 log_writer_read_output();
1697 sub log_writer_finish()
1699 return unless $log_pipe_pid;
1701 close($log_pipe_in);
1703 my $logger_failed = 0;
1704 my $read_result = log_writer_read_output(120);
1705 if ($read_result == -1) {
1706 $logger_failed = -1;
1707 Log (undef, "timed out reading from 'arv-put'");
1708 } elsif ($read_result != 0) {
1709 $logger_failed = -2;
1710 Log(undef, "failed to read arv-put log manifest to EOF");
1713 waitpid($log_pipe_pid, 0);
1715 $logger_failed ||= $?;
1716 Log(undef, "log_writer_finish: arv-put exited " . exit_status_s($?))
1719 close($log_pipe_out);
1720 my $arv_put_output = $logger_failed ? undef : $log_pipe_out_buf;
1721 $log_pipe_pid = $log_pipe_in = $log_pipe_out = $log_pipe_out_buf =
1722 $log_pipe_out_select = undef;
1724 return $arv_put_output;
1727 sub log_writer_is_active() {
1728 return $log_pipe_pid;
1731 sub Log # ($jobstepidx, $logmessage)
1733 my ($jobstepidx, $logmessage) = @_;
1734 if ($logmessage =~ /\n/) {
1735 for my $line (split (/\n/, $_[1])) {
1736 Log ($jobstepidx, $line);
1740 my $fh = select STDERR; $|=1; select $fh;
1742 if (defined($jobstepidx) && exists($jobstep[$jobstepidx]->{arvados_task})) {
1743 $task_qseq = $jobstepidx;
1745 my $message = sprintf ("%s %d %s %s", $job_id, $$, $task_qseq, $logmessage);
1746 $message =~ s{([^ -\176])}{"\\" . sprintf ("%03o", ord($1))}ge;
1749 if (log_writer_is_active() || -t STDERR) {
1750 my @gmtime = gmtime;
1751 $datetime = sprintf ("%04d-%02d-%02d_%02d:%02d:%02d",
1752 $gmtime[5]+1900, $gmtime[4]+1, @gmtime[3,2,1,0]);
1754 print STDERR ((-t STDERR) ? ($datetime." ".$message) : $message);
1756 if (log_writer_is_active()) {
1757 log_writer_send($datetime . " " . $message);
1764 my ($package, $file, $line) = caller;
1765 my $message = "@_ at $file line $line\n";
1766 Log (undef, $message);
1767 freeze() if @jobstep_todo;
1768 create_output_collection() if @jobstep_todo;
1778 if ($Job->{'state'} eq 'Cancelled') {
1779 $Job->update_attributes('finished_at' => scalar gmtime);
1781 $Job->update_attributes('state' => 'Failed');
1788 my $justcheckpoint = shift; # false if this will be the last meta saved
1789 return if $justcheckpoint; # checkpointing is not relevant post-Warehouse.pm
1790 return unless log_writer_is_active();
1791 my $log_manifest = log_writer_finish();
1792 return unless defined($log_manifest);
1795 my $prev_log_coll = api_call("collections/get", uuid => $Job->{log});
1796 $log_manifest = $prev_log_coll->{manifest_text} . $log_manifest;
1799 my $log_coll = api_call(
1800 "collections/create", ensure_unique_name => 1, collection => {
1801 manifest_text => $log_manifest,
1802 owner_uuid => $Job->{owner_uuid},
1803 name => sprintf("Log from %s job %s", $Job->{script}, $Job->{uuid}),
1805 Log(undef, "log collection is " . $log_coll->{portable_data_hash});
1806 $Job->update_attributes('log' => $log_coll->{portable_data_hash});
1810 sub freeze_if_want_freeze
1812 if ($main::please_freeze)
1814 release_allocation();
1817 # kill some srun procs before freeze+stop
1818 map { $proc{$_} = {} } @_;
1821 killem (keys %proc);
1822 select (undef, undef, undef, 0.1);
1824 while (($died = waitpid (-1, WNOHANG)) > 0)
1826 delete $proc{$died};
1831 create_output_collection();
1841 Log (undef, "Freeze not implemented");
1848 croak ("Thaw not implemented");
1864 $s =~ s{\\(.)}{$1 eq "n" ? "\n" : $1}ge;
1871 my $srunargs = shift;
1872 my $execargs = shift;
1873 my $opts = shift || {};
1876 my $label = exists $opts->{label} ? $opts->{label} : "@$execargs";
1877 Log (undef, "$label: start");
1879 my ($stderr_r, $stderr_w);
1880 pipe $stderr_r, $stderr_w or croak("pipe() failed: $!");
1882 my ($stdout_r, $stdout_w);
1883 pipe $stdout_r, $stdout_w or croak("pipe() failed: $!");
1885 my $srunpid = fork();
1890 fcntl($stderr_w, F_SETFL, 0) or croak($!); # no close-on-exec
1891 fcntl($stdout_w, F_SETFL, 0) or croak($!);
1892 open(STDERR, ">&", $stderr_w);
1893 open(STDOUT, ">&", $stdout_w);
1894 srun ($srunargs, $execargs, $opts, $stdin);
1900 set_nonblocking($stderr_r);
1901 set_nonblocking($stdout_r);
1903 # Add entries to @jobstep and %proc so check_squeue() and
1904 # freeze_if_want_freeze() can treat it like a job task process.
1908 stderr_captured => '',
1909 stdout_r => $stdout_r,
1910 stdout_captured => '',
1912 my $jobstepidx = $#jobstep;
1914 jobstepidx => $jobstepidx,
1916 $reader{$jobstepidx} = $stderr_r;
1918 while ($srunpid != waitpid ($srunpid, WNOHANG)) {
1919 my $busy = readfrompipes();
1920 if (!$busy || ($latest_refresh + 2 < scalar time)) {
1921 check_refresh_wanted();
1925 select(undef, undef, undef, 0.1);
1927 killem(keys %proc) if $main::please_freeze;
1931 1 while readfrompipes();
1932 process_stderr_final ($jobstepidx);
1934 Log (undef, "$label: exit ".exit_status_s($exited));
1938 delete $proc{$srunpid};
1939 delete $reader{$jobstepidx};
1941 my $j = pop @jobstep;
1942 return ($exited, $j->{stdout_captured}, $j->{stderr_captured});
1948 my $srunargs = shift;
1949 my $execargs = shift;
1950 my $opts = shift || {};
1952 my $args = $have_slurm ? [@$srunargs, @$execargs] : $execargs;
1954 $Data::Dumper::Terse = 1;
1955 $Data::Dumper::Indent = 0;
1956 my $show_cmd = Dumper($args);
1957 $show_cmd =~ s/(TOKEN\\*=)[^\s\']+/${1}[...]/g;
1958 $show_cmd =~ s/\n/ /g;
1959 if ($opts->{fork}) {
1960 Log(undef, "starting: $show_cmd");
1962 # This is a child process: parent is in charge of reading our
1963 # stderr and copying it to Log() if needed.
1964 warn "starting: $show_cmd\n";
1967 if (defined $stdin) {
1968 my $child = open STDIN, "-|";
1969 defined $child or die "no fork: $!";
1971 print $stdin or die $!;
1972 close STDOUT or die $!;
1977 return system (@$args) if $opts->{fork};
1980 warn "ENV size is ".length(join(" ",%ENV));
1981 die "exec failed: $!: @$args";
1985 sub ban_node_by_slot {
1986 # Don't start any new jobsteps on this node for 60 seconds
1988 $slot[$slotid]->{node}->{hold_until} = 60 + scalar time;
1989 $slot[$slotid]->{node}->{hold_count}++;
1990 Log (undef, "backing off node " . $slot[$slotid]->{node}->{name} . " for 60 seconds");
1995 my ($lockfile, $error_message) = @_;
1996 open L, ">", $lockfile or croak("$lockfile: $!");
1997 if (!flock L, LOCK_EX|LOCK_NB) {
1998 croak("Can't lock $lockfile: $error_message\n");
2002 sub find_docker_image {
2003 # Given a Keep locator, check to see if it contains a Docker image.
2004 # If so, return its stream name and Docker hash.
2005 # If not, return undef for both values.
2006 my $locator = shift;
2007 my ($streamname, $filename);
2008 my $image = api_call("collections/get", uuid => $locator);
2010 foreach my $line (split(/\n/, $image->{manifest_text})) {
2011 my @tokens = split(/\s+/, $line);
2013 $streamname = shift(@tokens);
2014 foreach my $filedata (grep(/^\d+:\d+:/, @tokens)) {
2015 if (defined($filename)) {
2016 return (undef, undef); # More than one file in the Collection.
2018 $filename = (split(/:/, $filedata, 3))[2];
2023 if (defined($filename) and ($filename =~ /^([0-9A-Fa-f]{64})\.tar$/)) {
2024 return ($streamname, $1);
2026 return (undef, undef);
2031 # Calculate the number of times an operation should be retried,
2032 # assuming exponential backoff, and that we're willing to retry as
2033 # long as tasks have been running. Enforce a minimum of 3 retries.
2034 my ($starttime, $endtime, $timediff, $retries);
2036 $starttime = $jobstep[0]->{starttime};
2037 $endtime = $jobstep[-1]->{finishtime};
2039 if (!defined($starttime)) {
2041 } elsif (!defined($endtime)) {
2042 $timediff = time - $starttime;
2044 $timediff = ($endtime - $starttime) - (time - $endtime);
2046 if ($timediff > 0) {
2047 $retries = int(log($timediff) / log(2));
2049 $retries = 1; # Use the minimum.
2051 return ($retries > 3) ? $retries : 3;
2055 # Pass in two function references.
2056 # This method will be called with the remaining arguments.
2057 # If it dies, retry it with exponential backoff until it succeeds,
2058 # or until the current retry_count is exhausted. After each failure
2059 # that can be retried, the second function will be called with
2060 # the current try count (0-based), next try time, and error message.
2061 my $operation = shift;
2062 my $retry_callback = shift;
2063 my $retries = retry_count();
2064 foreach my $try_count (0..$retries) {
2065 my $next_try = time + (2 ** $try_count);
2066 my $result = eval { $operation->(@_); };
2069 } elsif ($try_count < $retries) {
2070 $retry_callback->($try_count, $next_try, $@);
2071 my $sleep_time = $next_try - time;
2072 sleep($sleep_time) if ($sleep_time > 0);
2075 # Ensure the error message ends in a newline, so Perl doesn't add
2076 # retry_op's line number to it.
2082 # Pass in a /-separated API method name, and arguments for it.
2083 # This function will call that method, retrying as needed until
2084 # the current retry_count is exhausted, with a log on the first failure.
2085 my $method_name = shift;
2086 my $log_api_retry = sub {
2087 my ($try_count, $next_try_at, $errmsg) = @_;
2088 $errmsg =~ s/\s*\bat \Q$0\E line \d+\.?\s*//;
2089 $errmsg =~ s/\s/ /g;
2090 $errmsg =~ s/\s+$//;
2092 if ($next_try_at < time) {
2093 $retry_msg = "Retrying.";
2095 my $next_try_fmt = strftime "%Y-%m-%dT%H:%M:%SZ", gmtime($next_try_at);
2096 $retry_msg = "Retrying at $next_try_fmt.";
2098 Log(undef, "API method $method_name failed: $errmsg. $retry_msg");
2101 foreach my $key (split(/\//, $method_name)) {
2102 $method = $method->{$key};
2104 return retry_op(sub { $method->execute(@_); }, $log_api_retry, @_);
2108 # Given a $?, return a human-readable exit code string like "0" or
2109 # "1" or "0 with signal 1" or "1 with signal 11".
2110 my $exitcode = shift;
2111 my $s = $exitcode >> 8;
2112 if ($exitcode & 0x7f) {
2113 $s .= " with signal " . ($exitcode & 0x7f);
2115 if ($exitcode & 0x80) {
2116 $s .= " with core dump";
2121 sub handle_readall {
2122 # Pass in a glob reference to a file handle.
2123 # Read all its contents and return them as a string.
2124 my $fh_glob_ref = shift;
2126 return <$fh_glob_ref>;
2129 sub tar_filename_n {
2131 return sprintf("%s/git.%s.%d.tar", $ENV{CRUNCH_TMP}, $job_id, $n);
2134 sub add_git_archive {
2135 # Pass in a git archive command as a string or list, a la system().
2136 # This method will save its output to be included in the archive sent to the
2140 if (!open(GIT_ARCHIVE, ">", tar_filename_n($git_tar_count))) {
2141 croak("Failed to save git archive: $!");
2143 my $git_pid = open2(">&GIT_ARCHIVE", $git_input, @_);
2145 waitpid($git_pid, 0);
2148 croak("Failed to save git archive: git exited " . exit_status_s($?));
2152 sub combined_git_archive {
2153 # Combine all saved tar archives into a single archive, then return its
2154 # contents in a string. Return undef if no archives have been saved.
2155 if ($git_tar_count < 1) {
2158 my $base_tar_name = tar_filename_n(1);
2159 foreach my $tar_to_append (map { tar_filename_n($_); } (2..$git_tar_count)) {
2160 my $tar_exit = system("tar", "-Af", $base_tar_name, $tar_to_append);
2161 if ($tar_exit != 0) {
2162 croak("Error preparing build archive: tar -A exited " .
2163 exit_status_s($tar_exit));
2166 if (!open(GIT_TAR, "<", $base_tar_name)) {
2167 croak("Could not open build archive: $!");
2169 my $tar_contents = handle_readall(\*GIT_TAR);
2171 return $tar_contents;
2174 sub set_nonblocking {
2176 my $flags = fcntl ($fh, F_GETFL, 0) or croak ($!);
2177 fcntl ($fh, F_SETFL, $flags | O_NONBLOCK) or croak ($!);
2183 # This is crunch-job's internal dispatch script. crunch-job running on the API
2184 # server invokes this script on individual compute nodes, or localhost if we're
2185 # running a job locally. It gets called in two modes:
2187 # * No arguments: Installation mode. Read a tar archive from the DATA
2188 # file handle; it includes the Crunch script's source code, and
2189 # maybe SDKs as well. Those should be installed in the proper
2190 # locations. This runs outside of any Docker container, so don't try to
2191 # introspect Crunch's runtime environment.
2193 # * With arguments: Crunch script run mode. This script should set up the
2194 # environment, then run the command specified in the arguments. This runs
2195 # inside any Docker container.
2198 use File::Path qw( make_path remove_tree );
2199 use POSIX qw(getcwd);
2201 use constant TASK_TEMPFAIL => 111;
2203 # Map SDK subdirectories to the path environments they belong to.
2204 my %SDK_ENVVARS = ("perl/lib" => "PERLLIB", "ruby/lib" => "RUBYLIB");
2206 my $destdir = $ENV{"CRUNCH_SRC"};
2207 my $archive_hash = $ENV{"CRUNCH_GIT_ARCHIVE_HASH"};
2208 my $repo = $ENV{"CRUNCH_SRC_URL"};
2209 my $install_dir = $ENV{"CRUNCH_INSTALL"} || (getcwd() . "/opt");
2210 my $job_work = $ENV{"JOB_WORK"};
2211 my $task_work = $ENV{"TASK_WORK"};
2213 open(STDOUT_ORIG, ">&", STDOUT);
2214 open(STDERR_ORIG, ">&", STDERR);
2216 for my $dir ($destdir, $job_work, $task_work) {
2219 -e $dir or die "Failed to create temporary directory ($dir): $!";
2224 remove_tree($task_work, {keep_root => 1});
2227 ### Crunch script run mode
2229 # We want to do routine logging during task 0 only. This gives the user
2230 # the information they need, but avoids repeating the information for every
2233 if ($ENV{TASK_SEQUENCE} eq "0") {
2236 printf STDERR_ORIG "[Crunch] $msg\n", @_;
2242 my $python_src = "$install_dir/python";
2243 my $venv_dir = "$job_work/.arvados.venv";
2244 my $venv_built = -e "$venv_dir/bin/activate";
2245 if ((!$venv_built) and (-d $python_src) and can_run("virtualenv")) {
2246 shell_or_die(undef, "virtualenv", "--quiet", "--system-site-packages",
2247 "--python=python2.7", $venv_dir);
2248 shell_or_die(TASK_TEMPFAIL, "$venv_dir/bin/pip", "--quiet", "install", "-I", $python_src);
2250 $Log->("Built Python SDK virtualenv");
2253 my @pysdk_version_cmd = ("python", "-c",
2254 "from pkg_resources import get_distribution as get; print get('arvados-python-client').version");
2256 $Log->("Running in Python SDK virtualenv");
2257 @pysdk_version_cmd = ();
2258 my $orig_argv = join(" ", map { quotemeta($_); } @ARGV);
2259 @ARGV = ("/bin/sh", "-ec",
2260 ". \Q$venv_dir/bin/activate\E; exec $orig_argv");
2261 } elsif (-d $python_src) {
2262 $Log->("Warning: virtualenv not found inside Docker container default " .
2263 "\$PATH. Can't install Python SDK.");
2266 if (@pysdk_version_cmd) {
2267 open(my $pysdk_version_pipe, "-|", @pysdk_version_cmd);
2268 my $pysdk_version = <$pysdk_version_pipe>;
2269 close($pysdk_version_pipe);
2271 chomp($pysdk_version);
2272 $Log->("Using Arvados SDK version $pysdk_version");
2274 # A lot could've gone wrong here, but pretty much all of it means that
2275 # Python won't be able to load the Arvados SDK.
2276 $Log->("Warning: Arvados SDK not found");
2280 while (my ($sdk_dir, $sdk_envkey) = each(%SDK_ENVVARS)) {
2281 my $sdk_path = "$install_dir/$sdk_dir";
2283 if ($ENV{$sdk_envkey}) {
2284 $ENV{$sdk_envkey} = "$sdk_path:" . $ENV{$sdk_envkey};
2286 $ENV{$sdk_envkey} = $sdk_path;
2288 $Log->("Arvados SDK added to %s", $sdk_envkey);
2293 die "Cannot exec `@ARGV`: $!";
2296 ### Installation mode
2297 open L, ">", "$destdir.lock" or die "$destdir.lock: $!";
2299 if (readlink ("$destdir.archive_hash") eq $archive_hash && -d $destdir) {
2300 # This exact git archive (source + arvados sdk) is already installed
2301 # here, so there's no need to reinstall it.
2303 # We must consume our DATA section, though: otherwise the process
2304 # feeding it to us will get SIGPIPE.
2306 while (read(DATA, $buf, 65536)) { }
2311 unlink "$destdir.archive_hash";
2315 # Ignore SIGPIPE: we check retval of close() instead. See perlipc(1).
2316 local $SIG{PIPE} = "IGNORE";
2317 warn "Extracting archive: $archive_hash\n";
2318 # --ignore-zeros is necessary sometimes: depending on how much NUL
2319 # padding tar -A put on our combined archive (which in turn depends
2320 # on the length of the component archives) tar without
2321 # --ignore-zeros will exit before consuming stdin and cause close()
2322 # to fail on the resulting SIGPIPE.
2323 if (!open(TARX, "|-", "tar", "--ignore-zeros", "-xC", $destdir)) {
2324 die "Error launching 'tar -xC $destdir': $!";
2326 # If we send too much data to tar in one write (> 4-5 MiB), it stops, and we
2327 # get SIGPIPE. We must feed it data incrementally.
2329 while (read(DATA, $tar_input, 65536)) {
2330 print TARX $tar_input;
2333 die "'tar -xC $destdir' exited $?: $!";
2339 my $sdk_root = "$destdir/.arvados.sdk/sdk";
2341 foreach my $sdk_lang (("python",
2342 map { (split /\//, $_, 2)[0]; } keys(%SDK_ENVVARS))) {
2343 if (-d "$sdk_root/$sdk_lang") {
2344 if (!rename("$sdk_root/$sdk_lang", "$install_dir/$sdk_lang")) {
2345 die "Failed to install $sdk_lang SDK: $!";
2351 my $python_dir = "$install_dir/python";
2352 if ((-d $python_dir) and can_run("python2.7")) {
2353 open(my $egg_info_pipe, "-|",
2354 "python2.7 \Q$python_dir/setup.py\E egg_info 2>&1 >/dev/null");
2355 my @egg_info_errors = <$egg_info_pipe>;
2356 close($egg_info_pipe);
2359 if (@egg_info_errors and (($egg_info_errors[-1] =~ /\bgit\b/) or ($egg_info_errors[-1] =~ /\[Errno 2\]/))) {
2360 # egg_info apparently failed because it couldn't ask git for a build tag.
2361 # Specify no build tag.
2362 open(my $pysdk_cfg, ">>", "$python_dir/setup.cfg");
2363 print $pysdk_cfg "\n[egg_info]\ntag_build =\n";
2366 my $egg_info_exit = $? >> 8;
2367 foreach my $errline (@egg_info_errors) {
2370 warn "python setup.py egg_info failed: exit $egg_info_exit";
2371 exit ($egg_info_exit || 1);
2376 # Hide messages from the install script (unless it fails: shell_or_die
2377 # will show $destdir.log in that case).
2378 open(STDOUT, ">>", "$destdir.log");
2379 open(STDERR, ">&", STDOUT);
2381 if (-e "$destdir/crunch_scripts/install") {
2382 shell_or_die (undef, "$destdir/crunch_scripts/install", $install_dir);
2383 } elsif (!-e "./install.sh" && -e "./tests/autotests.sh") {
2385 shell_or_die (undef, "./tests/autotests.sh", $install_dir);
2386 } elsif (-e "./install.sh") {
2387 shell_or_die (undef, "./install.sh", $install_dir);
2390 if ($archive_hash) {
2391 unlink "$destdir.archive_hash.new";
2392 symlink ($archive_hash, "$destdir.archive_hash.new") or die "$destdir.archive_hash.new: $!";
2393 rename ("$destdir.archive_hash.new", "$destdir.archive_hash") or die "$destdir.archive_hash: $!";
2399 my $command_name = shift;
2400 open(my $which, "-|", "which", $command_name);
2401 while (<$which>) { }
2408 my $exitcode = shift;
2410 if ($ENV{"DEBUG"}) {
2411 print STDERR "@_\n";
2413 if (system (@_) != 0) {
2416 my $exitstatus = sprintf("exit %d signal %d", $code >> 8, $code & 0x7f);
2417 open STDERR, ">&STDERR_ORIG";
2418 system ("cat $destdir.log >&2");
2419 warn "@_ failed ($err): $exitstatus";
2420 if (defined($exitcode)) {
2424 exit (($code >> 8) || 1);