Coverage for python/lsst/ctrl/bps/htcondor/lssthtc.py: 13%
605 statements
« prev ^ index » next coverage.py v7.4.4, created at 2024-04-20 02:53 -0700
1# This file is part of ctrl_bps_htcondor.
2#
3# Developed for the LSST Data Management System.
4# This product includes software developed by the LSST Project
5# (https://www.lsst.org).
6# See the COPYRIGHT file at the top-level directory of this distribution
7# for details of code ownership.
8#
9# This software is dual licensed under the GNU General Public License and also
10# under a 3-clause BSD license. Recipients may choose which of these licenses
11# to use; please see the files gpl-3.0.txt and/or bsd_license.txt,
12# respectively. If you choose the GPL option then the following text applies
13# (but note that there is still no warranty even if you opt for BSD instead):
14#
15# This program is free software: you can redistribute it and/or modify
16# it under the terms of the GNU General Public License as published by
17# the Free Software Foundation, either version 3 of the License, or
18# (at your option) any later version.
19#
20# This program is distributed in the hope that it will be useful,
21# but WITHOUT ANY WARRANTY; without even the implied warranty of
22# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
23# GNU General Public License for more details.
24#
25# You should have received a copy of the GNU General Public License
26# along with this program. If not, see <https://www.gnu.org/licenses/>.
28"""Placeholder HTCondor DAGMan API.
30There is new work on a python DAGMan API from HTCondor. However, at this
31time, it tries to make things easier by assuming DAG is easily broken into
32levels where there are 1-1 or all-to-all relationships to nodes in next
33level. LSST workflows are more complicated.
34"""
36__all__ = [
37 "DagStatus",
38 "JobStatus",
39 "NodeStatus",
40 "RestrictedDict",
41 "HTCJob",
42 "HTCDag",
43 "htc_backup_files",
44 "htc_check_dagman_output",
45 "htc_create_submit_from_cmd",
46 "htc_create_submit_from_dag",
47 "htc_create_submit_from_file",
48 "htc_escape",
49 "htc_write_attribs",
50 "htc_write_condor_file",
51 "htc_query_history",
52 "htc_query_present",
53 "htc_version",
54 "htc_submit_dag",
55 "condor_history",
56 "condor_q",
57 "condor_search",
58 "condor_status",
59 "update_job_info",
60 "MISSING_ID",
61 "summary_from_dag",
62 "read_dag_info",
63 "read_dag_log",
64 "read_dag_nodes_log",
65 "read_dag_status",
66 "read_node_status",
67 "write_dag_info",
68 "pegasus_name_to_label",
69]
72import itertools
73import json
74import logging
75import os
76import pprint
77import re
78import subprocess
79from collections import defaultdict
80from collections.abc import MutableMapping
81from datetime import datetime, timedelta
82from enum import IntEnum
83from pathlib import Path
85import classad
86import htcondor
87import networkx
88from packaging import version
90from .handlers import HTC_JOB_AD_HANDLERS
92_LOG = logging.getLogger(__name__)
94MISSING_ID = -99999
97class DagStatus(IntEnum):
98 """HTCondor DAGMan's statuses for a DAG."""
100 OK = 0
101 ERROR = 1 # an error condition different than those listed here
102 FAILED = 2 # one or more nodes in the DAG have failed
103 ABORTED = 3 # the DAG has been aborted by an ABORT-DAG-ON specification
104 REMOVED = 4 # the DAG has been removed by condor_rm
105 CYCLE = 5 # a cycle was found in the DAG
106 SUSPENDED = 6 # the DAG has been suspended (see section 2.10.8)
109class JobStatus(IntEnum):
110 """HTCondor's statuses for jobs."""
112 UNEXPANDED = 0 # Unexpanded
113 IDLE = 1 # Idle
114 RUNNING = 2 # Running
115 REMOVED = 3 # Removed
116 COMPLETED = 4 # Completed
117 HELD = 5 # Held
118 TRANSFERRING_OUTPUT = 6 # Transferring_Output
119 SUSPENDED = 7 # Suspended
122class NodeStatus(IntEnum):
123 """HTCondor's statuses for DAGman nodes."""
125 # (STATUS_NOT_READY): At least one parent has not yet finished or the node
126 # is a FINAL node.
127 NOT_READY = 0
129 # (STATUS_READY): All parents have finished, but the node is not yet
130 # running.
131 READY = 1
133 # (STATUS_PRERUN): The node’s PRE script is running.
134 PRERUN = 2
136 # (STATUS_SUBMITTED): The node’s HTCondor job(s) are in the queue.
137 # StatusDetails = "not_idle" -> running.
138 # JobProcsHeld = 1-> hold.
139 # JobProcsQueued = 1 -> idle.
140 SUBMITTED = 3
142 # (STATUS_POSTRUN): The node’s POST script is running.
143 POSTRUN = 4
145 # (STATUS_DONE): The node has completed successfully.
146 DONE = 5
148 # (STATUS_ERROR): The node has failed. StatusDetails has info (e.g.,
149 # ULOG_JOB_ABORTED for deleted job).
150 ERROR = 6
152 # (STATUS_FUTILE): The node will never run because ancestor node failed.
153 FUTILE = 7
156HTC_QUOTE_KEYS = {"environment"}
157HTC_VALID_JOB_KEYS = {
158 "universe",
159 "executable",
160 "arguments",
161 "environment",
162 "log",
163 "error",
164 "output",
165 "should_transfer_files",
166 "when_to_transfer_output",
167 "getenv",
168 "notification",
169 "notify_user",
170 "concurrency_limit",
171 "transfer_executable",
172 "transfer_input_files",
173 "transfer_output_files",
174 "request_cpus",
175 "request_memory",
176 "request_disk",
177 "priority",
178 "category",
179 "requirements",
180 "on_exit_hold",
181 "on_exit_hold_reason",
182 "on_exit_hold_subcode",
183 "max_retries",
184 "periodic_release",
185 "periodic_remove",
186 "accounting_group",
187 "accounting_group_user",
188}
189HTC_VALID_JOB_DAG_KEYS = {"vars", "pre", "post", "retry", "retry_unless_exit", "abort_dag_on", "abort_exit"}
190HTC_VERSION = version.parse(htcondor.__version__)
193class RestrictedDict(MutableMapping):
194 """A dictionary that only allows certain keys.
196 Parameters
197 ----------
198 valid_keys : `Container`
199 Strings that are valid keys.
200 init_data : `dict` or `RestrictedDict`, optional
201 Initial data.
203 Raises
204 ------
205 KeyError
206 If invalid key(s) in init_data.
207 """
209 def __init__(self, valid_keys, init_data=()):
210 self.valid_keys = valid_keys
211 self.data = {}
212 self.update(init_data)
214 def __getitem__(self, key):
215 """Return value for given key if exists.
217 Parameters
218 ----------
219 key : `str`
220 Identifier for value to return.
222 Returns
223 -------
224 value : `~collections.abc.Any`
225 Value associated with given key.
227 Raises
228 ------
229 KeyError
230 If key doesn't exist.
231 """
232 return self.data[key]
234 def __delitem__(self, key):
235 """Delete value for given key if exists.
237 Parameters
238 ----------
239 key : `str`
240 Identifier for value to delete.
242 Raises
243 ------
244 KeyError
245 If key doesn't exist.
246 """
247 del self.data[key]
249 def __setitem__(self, key, value):
250 """Store key,value in internal dict only if key is valid.
252 Parameters
253 ----------
254 key : `str`
255 Identifier to associate with given value.
256 value : `~collections.abc.Any`
257 Value to store.
259 Raises
260 ------
261 KeyError
262 If key is invalid.
263 """
264 if key not in self.valid_keys:
265 raise KeyError(f"Invalid key {key}")
266 self.data[key] = value
268 def __iter__(self):
269 return self.data.__iter__()
271 def __len__(self):
272 return len(self.data)
274 def __str__(self):
275 return str(self.data)
278def htc_backup_files(wms_path, subdir=None, limit=100):
279 """Backup select HTCondor files in the submit directory.
281 Files will be saved in separate subdirectories which will be created in
282 the submit directory where the files are located. These subdirectories
283 will be consecutive, zero-padded integers. Their values will correspond to
284 the number of HTCondor rescue DAGs in the submit directory.
286 Hence, with the default settings, copies after the initial failed run will
287 be placed in '001' subdirectory, '002' after the first restart, and so on
288 until the limit of backups is reached. If there's no rescue DAG yet, files
289 will be copied to '000' subdirectory.
291 Parameters
292 ----------
293 wms_path : `str` or `pathlib.Path`
294 Path to the submit directory either absolute or relative.
295 subdir : `str` or `pathlib.Path`, optional
296 A path, relative to the submit directory, where all subdirectories with
297 backup files will be kept. Defaults to None which means that the backup
298 subdirectories will be placed directly in the submit directory.
299 limit : `int`, optional
300 Maximal number of backups. If the number of backups reaches the limit,
301 the last backup files will be overwritten. The default value is 100
302 to match the default value of HTCondor's DAGMAN_MAX_RESCUE_NUM in
303 version 8.8+.
305 Raises
306 ------
307 FileNotFoundError
308 If the submit directory or the file that needs to be backed up does not
309 exist.
310 OSError
311 If the submit directory cannot be accessed or backing up a file failed
312 either due to permission or filesystem related issues.
314 Notes
315 -----
316 This is not a generic function for making backups. It is intended to be
317 used once, just before a restart, to make snapshots of files which will be
318 overwritten by HTCondor after during the next run.
319 """
320 width = len(str(limit))
322 path = Path(wms_path).resolve()
323 if not path.is_dir():
324 raise FileNotFoundError(f"Directory {path} not found")
326 # Initialize the backup counter.
327 rescue_dags = list(Path(wms_path).glob("*.rescue*"))
328 counter = min(len(rescue_dags), limit)
330 # Create the backup directory and move select files there.
331 dest = Path(wms_path)
332 if subdir:
333 # PurePath.is_relative_to() is not available before Python 3.9. Hence
334 # we need to check is 'subdir' is in the submit directory in some other
335 # way if it is an absolute path.
336 subdir = Path(subdir)
337 if subdir.is_absolute():
338 if dest not in subdir.parents:
339 _LOG.warning(
340 "Invalid backup location: '%s' not in the submit directory, will use '%s' instead.",
341 subdir,
342 wms_path,
343 )
344 else:
345 dest /= subdir
346 else:
347 dest /= subdir
348 dest /= f"{counter:0{width}}"
349 try:
350 dest.mkdir(parents=True, exist_ok=False if counter < limit else True)
351 except FileExistsError:
352 _LOG.warning("Refusing to do backups: target directory '%s' already exists", dest)
353 else:
354 for patt in ["*.info.*", "*.dag.metrics", "*.dag.nodes.log", "*.node_status"]:
355 for source in path.glob(patt):
356 if source.is_file():
357 target = dest / source.relative_to(path)
358 try:
359 source.rename(target)
360 except OSError as exc:
361 raise type(exc)(f"Backing up '{source}' failed: {exc.strerror}") from None
362 else:
363 raise FileNotFoundError(f"Backing up '{source}' failed: not a file")
366def htc_escape(value):
367 """Escape characters in given value based upon HTCondor syntax.
369 Parameters
370 ----------
371 value : `~collections.abc.Any`
372 Value that needs to have characters escaped if string.
374 Returns
375 -------
376 new_value : `~collections.abc.Any`
377 Given value with characters escaped appropriate for HTCondor if string.
378 """
379 if isinstance(value, str):
380 newval = value.replace('"', '""').replace("'", "''").replace(""", '"')
381 else:
382 newval = value
384 return newval
387def htc_write_attribs(stream, attrs):
388 """Write job attributes in HTCondor format to writeable stream.
390 Parameters
391 ----------
392 stream : `~io.TextIOBase`
393 Output text stream (typically an open file).
394 attrs : `dict`
395 HTCondor job attributes (dictionary of attribute key, value).
396 """
397 for key, value in attrs.items():
398 # Make sure strings are syntactically correct for HTCondor.
399 if isinstance(value, str):
400 pval = f'"{htc_escape(value)}"'
401 else:
402 pval = value
404 print(f"+{key} = {pval}", file=stream)
407def htc_write_condor_file(filename, job_name, job, job_attrs):
408 """Write an HTCondor submit file.
410 Parameters
411 ----------
412 filename : `str`
413 Filename for the HTCondor submit file.
414 job_name : `str`
415 Job name to use in submit file.
416 job : `RestrictedDict`
417 Submit script information.
418 job_attrs : `dict`
419 Job attributes.
420 """
421 os.makedirs(os.path.dirname(filename), exist_ok=True)
422 with open(filename, "w") as fh:
423 for key, value in job.items():
424 if value is not None:
425 if key in HTC_QUOTE_KEYS:
426 print(f'{key}="{htc_escape(value)}"', file=fh)
427 else:
428 print(f"{key}={value}", file=fh)
429 for key in ["output", "error", "log"]:
430 if key not in job:
431 filename = f"{job_name}.$(Cluster).${key[:3]}"
432 print(f"{key}={filename}", file=fh)
434 if job_attrs is not None:
435 htc_write_attribs(fh, job_attrs)
436 print("queue", file=fh)
439# To avoid doing the version check during every function call select
440# appropriate conversion function at the import time.
441#
442# Make sure that *each* version specific variant of the conversion function(s)
443# has the same signature after applying any changes!
444if HTC_VERSION < version.parse("8.9.8"): 444 ↛ 446line 444 didn't jump to line 446, because the condition on line 444 was never true
446 def htc_tune_schedd_args(**kwargs):
447 """Ensure that arguments for Schedd are version appropriate.
449 The old arguments: 'requirements' and 'attr_list' of
450 'Schedd.history()', 'Schedd.query()', and 'Schedd.xquery()' were
451 deprecated in favor of 'constraint' and 'projection', respectively,
452 starting from version 8.9.8. The function will convert "new" keyword
453 arguments to "old" ones.
455 Parameters
456 ----------
457 **kwargs
458 Any keyword arguments that Schedd.history(), Schedd.query(), and
459 Schedd.xquery() accepts.
461 Returns
462 -------
463 kwargs : `dict` [`str`, Any]
464 Keywords arguments that are guaranteed to work with the Python
465 HTCondor API.
467 Notes
468 -----
469 Function doesn't validate provided keyword arguments beyond converting
470 selected arguments to their version specific form. For example,
471 it won't remove keywords that are not supported by the methods
472 mentioned earlier.
473 """
474 translation_table = {
475 "constraint": "requirements",
476 "projection": "attr_list",
477 }
478 for new, old in translation_table.items():
479 try:
480 kwargs[old] = kwargs.pop(new)
481 except KeyError:
482 pass
483 return kwargs
485else:
487 def htc_tune_schedd_args(**kwargs):
488 """Ensure that arguments for Schedd are version appropriate.
490 This is the fallback function if no version specific alteration are
491 necessary. Effectively, a no-op.
493 Parameters
494 ----------
495 **kwargs
496 Any keyword arguments that Schedd.history(), Schedd.query(), and
497 Schedd.xquery() accepts.
499 Returns
500 -------
501 kwargs : `dict` [`str`, Any]
502 Keywords arguments that were passed to the function.
503 """
504 return kwargs
507def htc_query_history(schedds, **kwargs):
508 """Fetch history records from the condor_schedd daemon.
510 Parameters
511 ----------
512 schedds : `htcondor.Schedd`
513 HTCondor schedulers which to query for job information.
514 **kwargs
515 Any keyword arguments that Schedd.history() accepts.
517 Yields
518 ------
519 schedd_name : `str`
520 Name of the HTCondor scheduler managing the job queue.
521 job_ad : `dict` [`str`, Any]
522 A dictionary representing HTCondor ClassAd describing a job. It maps
523 job attributes names to values of the ClassAd expressions they
524 represent.
525 """
526 # If not set, provide defaults for positional arguments.
527 kwargs.setdefault("constraint", None)
528 kwargs.setdefault("projection", [])
529 kwargs = htc_tune_schedd_args(**kwargs)
530 for schedd_name, schedd in schedds.items():
531 for job_ad in schedd.history(**kwargs):
532 yield schedd_name, dict(job_ad)
535def htc_query_present(schedds, **kwargs):
536 """Query the condor_schedd daemon for job ads.
538 Parameters
539 ----------
540 schedds : `htcondor.Schedd`
541 HTCondor schedulers which to query for job information.
542 **kwargs
543 Any keyword arguments that Schedd.xquery() accepts.
545 Yields
546 ------
547 schedd_name : `str`
548 Name of the HTCondor scheduler managing the job queue.
549 job_ad : `dict` [`str`, Any]
550 A dictionary representing HTCondor ClassAd describing a job. It maps
551 job attributes names to values of the ClassAd expressions they
552 represent.
553 """
554 kwargs = htc_tune_schedd_args(**kwargs)
555 for schedd_name, schedd in schedds.items():
556 for job_ad in schedd.query(**kwargs):
557 yield schedd_name, dict(job_ad)
560def htc_version():
561 """Return the version given by the HTCondor API.
563 Returns
564 -------
565 version : `str`
566 HTCondor version as easily comparable string.
567 """
568 return str(HTC_VERSION)
571def htc_submit_dag(sub):
572 """Submit job for execution.
574 Parameters
575 ----------
576 sub : `htcondor.Submit`
577 An object representing a job submit description.
579 Returns
580 -------
581 schedd_job_info : `dict` [`str`, `dict` [`str`, `dict` [`str` Any]]]
582 Information about jobs satisfying the search criteria where for each
583 Scheduler, local HTCondor job ids are mapped to their respective
584 classads.
585 """
586 coll = htcondor.Collector()
587 schedd_ad = coll.locate(htcondor.DaemonTypes.Schedd)
588 schedd = htcondor.Schedd(schedd_ad)
590 # If Schedd.submit() fails, the method will raise an exception. Usually,
591 # that implies issues with the HTCondor pool which BPS can't address.
592 # Hence, no effort is made to handle the exception.
593 submit_result = schedd.submit(sub)
595 # Sadly, the ClassAd from Schedd.submit() (see above) does not have
596 # 'GlobalJobId' so we need to run a regular query to get it anyway.
597 schedd_name = schedd_ad["Name"]
598 schedd_dag_info = condor_q(
599 constraint=f"ClusterId == {submit_result.cluster()}", schedds={schedd_name: schedd}
600 )
601 return schedd_dag_info
604def htc_create_submit_from_dag(dag_filename, submit_options=None):
605 """Create a DAGMan job submit description.
607 Parameters
608 ----------
609 dag_filename : `str`
610 Name of file containing HTCondor DAG commands.
611 submit_options : `dict` [`str`, Any], optional
612 Contains extra options for command line (Value of None means flag).
614 Returns
615 -------
616 sub : `htcondor.Submit`
617 An object representing a job submit description.
619 Notes
620 -----
621 Use with HTCondor versions which support htcondor.Submit.from_dag(),
622 i.e., 8.9.3 or newer.
623 """
624 return htcondor.Submit.from_dag(dag_filename, submit_options)
627def htc_create_submit_from_cmd(dag_filename, submit_options=None):
628 """Create a DAGMan job submit description.
630 Create a DAGMan job submit description by calling ``condor_submit_dag``
631 on given DAG description file.
633 Parameters
634 ----------
635 dag_filename : `str`
636 Name of file containing HTCondor DAG commands.
637 submit_options : `dict` [`str`, Any], optional
638 Contains extra options for command line (Value of None means flag).
640 Returns
641 -------
642 sub : `htcondor.Submit`
643 An object representing a job submit description.
645 Notes
646 -----
647 Use with HTCondor versions which do not support htcondor.Submit.from_dag(),
648 i.e., older than 8.9.3.
649 """
650 # Run command line condor_submit_dag command.
651 cmd = "condor_submit_dag -f -no_submit -notification never -autorescue 1 -UseDagDir -no_recurse "
653 if submit_options is not None:
654 for opt, val in submit_options.items():
655 cmd += f" -{opt} {val or ''}"
656 cmd += f"{dag_filename}"
658 process = subprocess.Popen(
659 cmd.split(), shell=False, stdout=subprocess.PIPE, stderr=subprocess.STDOUT, encoding="utf-8"
660 )
661 process.wait()
663 if process.returncode != 0:
664 print(f"Exit code: {process.returncode}")
665 print(process.communicate()[0])
666 raise RuntimeError("Problems running condor_submit_dag")
668 return htc_create_submit_from_file(f"{dag_filename}.condor.sub")
671def htc_create_submit_from_file(submit_file):
672 """Parse a submission file.
674 Parameters
675 ----------
676 submit_file : `str`
677 Name of the HTCondor submit file.
679 Returns
680 -------
681 sub : `htcondor.Submit`
682 An object representing a job submit description.
683 """
684 descriptors = {}
685 with open(submit_file) as fh:
686 for line in fh:
687 line = line.strip()
688 if not line.startswith("#") and not line == "queue":
689 (key, val) = re.split(r"\s*=\s*", line, 1)
690 descriptors[key] = val
692 # Avoid UserWarning: the line 'copy_to_spool = False' was
693 # unused by Submit object. Is it a typo?
694 try:
695 del descriptors["copy_to_spool"]
696 except KeyError:
697 pass
699 return htcondor.Submit(descriptors)
702def _htc_write_job_commands(stream, name, jobs):
703 """Output the DAGMan job lines for single job in DAG.
705 Parameters
706 ----------
707 stream : `~io.TextIOBase`
708 Writeable text stream (typically an opened file).
709 name : `str`
710 Job name.
711 jobs : `RestrictedDict`
712 DAG job keys and values.
713 """
714 if "pre" in jobs:
715 print(
716 f"SCRIPT {jobs['pre'].get('defer', '')} PRE {name}"
717 f"{jobs['pre']['executable']} {jobs['pre'].get('arguments', '')}",
718 file=stream,
719 )
721 if "post" in jobs:
722 print(
723 f"SCRIPT {jobs['post'].get('defer', '')} PRE {name}"
724 f"{jobs['post']['executable']} {jobs['post'].get('arguments', '')}",
725 file=stream,
726 )
728 if "vars" in jobs:
729 for key, value in jobs["vars"]:
730 print(f'VARS {name} {key}="{htc_escape(value)}"', file=stream)
732 if "pre_skip" in jobs:
733 print(f"PRE_SKIP {name} {jobs['pre_skip']}", file=stream)
735 if "retry" in jobs and jobs["retry"]:
736 print(f"RETRY {name} {jobs['retry']} ", end="", file=stream)
737 if "retry_unless_exit" in jobs:
738 print(f"UNLESS-EXIT {jobs['retry_unless_exit']}", end="", file=stream)
739 print("\n", file=stream)
741 if "abort_dag_on" in jobs and jobs["abort_dag_on"]:
742 print(
743 f"ABORT-DAG-ON {name} {jobs['abort_dag_on']['node_exit']}"
744 f" RETURN {jobs['abort_dag_on']['abort_exit']}",
745 file=stream,
746 )
749class HTCJob:
750 """HTCondor job for use in building DAG.
752 Parameters
753 ----------
754 name : `str`
755 Name of the job.
756 label : `str`
757 Label that can used for grouping or lookup.
758 initcmds : `RestrictedDict`
759 Initial job commands for submit file.
760 initdagcmds : `RestrictedDict`
761 Initial commands for job inside DAG.
762 initattrs : `dict`
763 Initial dictionary of job attributes.
764 """
766 def __init__(self, name, label=None, initcmds=(), initdagcmds=(), initattrs=None):
767 self.name = name
768 self.label = label
769 self.cmds = RestrictedDict(HTC_VALID_JOB_KEYS, initcmds)
770 self.dagcmds = RestrictedDict(HTC_VALID_JOB_DAG_KEYS, initdagcmds)
771 self.attrs = initattrs
772 self.subfile = None
774 def __str__(self):
775 return self.name
777 def add_job_cmds(self, new_commands):
778 """Add commands to Job (overwrite existing).
780 Parameters
781 ----------
782 new_commands : `dict`
783 Submit file commands to be added to Job.
784 """
785 self.cmds.update(new_commands)
787 def add_dag_cmds(self, new_commands):
788 """Add DAG commands to Job (overwrite existing).
790 Parameters
791 ----------
792 new_commands : `dict`
793 DAG file commands to be added to Job.
794 """
795 self.dagcmds.update(new_commands)
797 def add_job_attrs(self, new_attrs):
798 """Add attributes to Job (overwrite existing).
800 Parameters
801 ----------
802 new_attrs : `dict`
803 Attributes to be added to Job.
804 """
805 if self.attrs is None:
806 self.attrs = {}
807 if new_attrs:
808 self.attrs.update(new_attrs)
810 def write_submit_file(self, submit_path, job_subdir=""):
811 """Write job description to submit file.
813 Parameters
814 ----------
815 submit_path : `str`
816 Prefix path for the submit file.
817 job_subdir : `str`, optional
818 Template for job subdir.
819 """
820 if not self.subfile:
821 self.subfile = f"{self.name}.sub"
822 job_subdir = job_subdir.format(self=self)
823 if job_subdir:
824 self.subfile = os.path.join(job_subdir, self.subfile)
825 htc_write_condor_file(os.path.join(submit_path, self.subfile), self.name, self.cmds, self.attrs)
827 def write_dag_commands(self, stream):
828 """Write DAG commands for single job to output stream.
830 Parameters
831 ----------
832 stream : `IO` or `str`
833 Output Stream.
834 """
835 print(f"JOB {self.name} {self.subfile}", file=stream)
836 _htc_write_job_commands(stream, self.name, self.dagcmds)
838 def dump(self, fh):
839 """Dump job information to output stream.
841 Parameters
842 ----------
843 fh : `~io.TextIOBase`
844 Output stream.
845 """
846 printer = pprint.PrettyPrinter(indent=4, stream=fh)
847 printer.pprint(self.name)
848 printer.pprint(self.cmds)
849 printer.pprint(self.attrs)
852class HTCDag(networkx.DiGraph):
853 """HTCondor DAG.
855 Parameters
856 ----------
857 data : networkx.DiGraph.data
858 Initial graph.
859 name : `str`
860 Name for DAG.
861 """
863 def __init__(self, data=None, name=""):
864 super().__init__(data=data, name=name)
866 self.graph["attr"] = {}
867 self.graph["run_id"] = None
868 self.graph["submit_path"] = None
869 self.graph["final_job"] = None
871 def __str__(self):
872 """Represent basic DAG info as string.
874 Returns
875 -------
876 info : `str`
877 String containing basic DAG info.
878 """
879 return f"{self.graph['name']} {len(self)}"
881 def add_attribs(self, attribs=None):
882 """Add attributes to the DAG.
884 Parameters
885 ----------
886 attribs : `dict`
887 DAG attributes.
888 """
889 if attribs is not None:
890 self.graph["attr"].update(attribs)
892 def add_job(self, job, parent_names=None, child_names=None):
893 """Add an HTCJob to the HTCDag.
895 Parameters
896 ----------
897 job : `HTCJob`
898 HTCJob to add to the HTCDag.
899 parent_names : `~collections.abc.Iterable` [`str`], optional
900 Names of parent jobs.
901 child_names : `~collections.abc.Iterable` [`str`], optional
902 Names of child jobs.
903 """
904 assert isinstance(job, HTCJob)
906 # Add dag level attributes to each job
907 job.add_job_attrs(self.graph["attr"])
909 self.add_node(job.name, data=job)
911 if parent_names is not None:
912 self.add_job_relationships(parent_names, job.name)
914 if child_names is not None:
915 self.add_job_relationships(child_names, job.name)
917 def add_job_relationships(self, parents, children):
918 """Add DAG edge between parents and children jobs.
920 Parameters
921 ----------
922 parents : `list` [`str`]
923 Contains parent job name(s).
924 children : `list` [`str`]
925 Contains children job name(s).
926 """
927 self.add_edges_from(itertools.product(parents, children))
929 def add_final_job(self, job):
930 """Add an HTCJob for the FINAL job in HTCDag.
932 Parameters
933 ----------
934 job : `HTCJob`
935 HTCJob to add to the HTCDag as a FINAL job.
936 """
937 # Add dag level attributes to each job
938 job.add_job_attrs(self.graph["attr"])
940 self.graph["final_job"] = job
942 def del_job(self, job_name):
943 """Delete the job from the DAG.
945 Parameters
946 ----------
947 job_name : `str`
948 Name of job in DAG to delete.
949 """
950 # Reconnect edges around node to delete
951 parents = self.predecessors(job_name)
952 children = self.successors(job_name)
953 self.add_edges_from(itertools.product(parents, children))
955 # Delete job node (which deletes its edges).
956 self.remove_node(job_name)
958 def write(self, submit_path, job_subdir=""):
959 """Write DAG to a file.
961 Parameters
962 ----------
963 submit_path : `str`
964 Prefix path for dag filename to be combined with DAG name.
965 job_subdir : `str`, optional
966 Template for job subdir.
967 """
968 self.graph["submit_path"] = submit_path
969 self.graph["dag_filename"] = os.path.join(submit_path, f"{self.graph['name']}.dag")
970 os.makedirs(submit_path, exist_ok=True)
971 with open(self.graph["dag_filename"], "w") as fh:
972 for _, nodeval in self.nodes().items():
973 job = nodeval["data"]
974 job.write_submit_file(submit_path, job_subdir)
975 job.write_dag_commands(fh)
976 for edge in self.edges():
977 print(f"PARENT {edge[0]} CHILD {edge[1]}", file=fh)
978 print(f"DOT {self.name}.dot", file=fh)
979 print(f"NODE_STATUS_FILE {self.name}.node_status", file=fh)
981 # Add bps attributes to dag submission
982 for key, value in self.graph["attr"].items():
983 print(f'SET_JOB_ATTR {key}= "{htc_escape(value)}"', file=fh)
985 if self.graph["final_job"]:
986 job = self.graph["final_job"]
987 job.write_submit_file(submit_path, job_subdir)
988 print(f"FINAL {job.name} {job.subfile}", file=fh)
989 if "pre" in job.dagcmds:
990 print(f"SCRIPT PRE {job.name} {job.dagcmds['pre']}", file=fh)
991 if "post" in job.dagcmds:
992 print(f"SCRIPT POST {job.name} {job.dagcmds['post']}", file=fh)
994 def dump(self, fh):
995 """Dump DAG info to output stream.
997 Parameters
998 ----------
999 fh : `io.IO` or `str`
1000 Where to dump DAG info as text.
1001 """
1002 for key, value in self.graph:
1003 print(f"{key}={value}", file=fh)
1004 for name, data in self.nodes().items():
1005 print(f"{name}:", file=fh)
1006 data.dump(fh)
1007 for edge in self.edges():
1008 print(f"PARENT {edge[0]} CHILD {edge[1]}", file=fh)
1009 if self.graph["final_job"]:
1010 print(f'FINAL {self.graph["final_job"].name}:', file=fh)
1011 self.graph["final_job"].dump(fh)
1013 def write_dot(self, filename):
1014 """Write a dot version of the DAG.
1016 Parameters
1017 ----------
1018 filename : `str`
1019 Name of the dot file.
1020 """
1021 pos = networkx.nx_agraph.graphviz_layout(self)
1022 networkx.draw(self, pos=pos)
1023 networkx.drawing.nx_pydot.write_dot(self, filename)
1026def condor_q(constraint=None, schedds=None, **kwargs):
1027 """Get information about the jobs in the HTCondor job queue(s).
1029 Parameters
1030 ----------
1031 constraint : `str`, optional
1032 Constraints to be passed to job query.
1033 schedds : `dict` [`str`, `htcondor.Schedd`], optional
1034 HTCondor schedulers which to query for job information. If None
1035 (default), the query will be run against local scheduler only.
1036 **kwargs : `~typing.Any`
1037 Additional keyword arguments that need to be passed to the internal
1038 query method.
1040 Returns
1041 -------
1042 job_info : `dict` [`str`, `dict` [`str`, `dict` [`str` Any]]]
1043 Information about jobs satisfying the search criteria where for each
1044 Scheduler, local HTCondor job ids are mapped to their respective
1045 classads.
1046 """
1047 return condor_query(constraint, schedds, htc_query_present, **kwargs)
1050def condor_history(constraint=None, schedds=None, **kwargs):
1051 """Get information about the jobs from HTCondor history records.
1053 Parameters
1054 ----------
1055 constraint : `str`, optional
1056 Constraints to be passed to job query.
1057 schedds : `dict` [`str`, `htcondor.Schedd`], optional
1058 HTCondor schedulers which to query for job information. If None
1059 (default), the query will be run against the history file of
1060 the local scheduler only.
1061 **kwargs : `~typing.Any`
1062 Additional keyword arguments that need to be passed to the internal
1063 query method.
1065 Returns
1066 -------
1067 job_info : `dict` [`str`, `dict` [`str`, `dict` [`str` Any]]]
1068 Information about jobs satisfying the search criteria where for each
1069 Scheduler, local HTCondor job ids are mapped to their respective
1070 classads.
1071 """
1072 return condor_query(constraint, schedds, htc_query_history, **kwargs)
1075def condor_query(constraint=None, schedds=None, query_func=htc_query_present, **kwargs):
1076 """Get information about HTCondor jobs.
1078 Parameters
1079 ----------
1080 constraint : `str`, optional
1081 Constraints to be passed to job query.
1082 schedds : `dict` [`str`, `htcondor.Schedd`], optional
1083 HTCondor schedulers which to query for job information. If None
1084 (default), the query will be run against the history file of
1085 the local scheduler only.
1086 query_func : callable
1087 An query function which takes following arguments:
1089 - ``schedds``: Schedulers to query (`list` [`htcondor.Schedd`]).
1090 - ``**kwargs``: Keyword arguments that will be passed to the query
1091 function.
1092 **kwargs : `~typing.Any`
1093 Additional keyword arguments that need to be passed to the query
1094 method.
1096 Returns
1097 -------
1098 job_info : `dict` [`str`, `dict` [`str`, `dict` [`str` Any]]]
1099 Information about jobs satisfying the search criteria where for each
1100 Scheduler, local HTCondor job ids are mapped to their respective
1101 classads.
1102 """
1103 if not schedds:
1104 coll = htcondor.Collector()
1105 schedd_ad = coll.locate(htcondor.DaemonTypes.Schedd)
1106 schedds = {schedd_ad["Name"]: htcondor.Schedd(schedd_ad)}
1108 # Make sure that 'ClusterId' and 'ProcId' attributes are always included
1109 # in the job classad. They are needed to construct the job id.
1110 added_attrs = set()
1111 if "projection" in kwargs and kwargs["projection"]:
1112 requested_attrs = set(kwargs["projection"])
1113 required_attrs = {"ClusterId", "ProcId"}
1114 added_attrs = required_attrs - requested_attrs
1115 for attr in added_attrs:
1116 kwargs["projection"].append(attr)
1118 unwanted_attrs = {"Env", "Environment"} | added_attrs
1119 job_info = defaultdict(dict)
1120 for schedd_name, job_ad in query_func(schedds, constraint=constraint, **kwargs):
1121 id_ = f"{job_ad['ClusterId']}.{job_ad['ProcId']}"
1122 for attr in set(job_ad) & unwanted_attrs:
1123 del job_ad[attr]
1124 job_info[schedd_name][id_] = job_ad
1125 _LOG.debug("query returned %d jobs", sum(len(val) for val in job_info.values()))
1127 # Restore the list of the requested attributes to its original value
1128 # if needed.
1129 if added_attrs:
1130 for attr in added_attrs:
1131 kwargs["projection"].remove(attr)
1133 # When returning the results filter out entries for schedulers with no jobs
1134 # matching the search criteria.
1135 return {key: val for key, val in job_info.items() if val}
1138def condor_search(constraint=None, hist=None, schedds=None):
1139 """Search for running and finished jobs satisfying given criteria.
1141 Parameters
1142 ----------
1143 constraint : `str`, optional
1144 Constraints to be passed to job query.
1145 hist : `float`
1146 Limit history search to this many days.
1147 schedds : `dict` [`str`, `htcondor.Schedd`], optional
1148 The list of the HTCondor schedulers which to query for job information.
1149 If None (default), only the local scheduler will be queried.
1151 Returns
1152 -------
1153 job_info : `dict` [`str`, `dict` [`str`, `dict` [`str` Any]]]
1154 Information about jobs satisfying the search criteria where for each
1155 Scheduler, local HTCondor job ids are mapped to their respective
1156 classads.
1157 """
1158 if not schedds:
1159 coll = htcondor.Collector()
1160 schedd_ad = coll.locate(htcondor.DaemonTypes.Schedd)
1161 schedds = {schedd_ad["Name"]: htcondor.Schedd(locate_ad=schedd_ad)}
1163 job_info = condor_q(constraint=constraint, schedds=schedds)
1164 if hist is not None:
1165 epoch = (datetime.now() - timedelta(days=hist)).timestamp()
1166 constraint += f" && (CompletionDate >= {epoch} || JobFinishedHookDone >= {epoch})"
1167 hist_info = condor_history(constraint, schedds=schedds)
1168 update_job_info(job_info, hist_info)
1169 return job_info
1172def condor_status(constraint=None, coll=None):
1173 """Get information about HTCondor pool.
1175 Parameters
1176 ----------
1177 constraint : `str`, optional
1178 Constraints to be passed to the query.
1179 coll : `htcondor.Collector`, optional
1180 Object representing HTCondor collector daemon.
1182 Returns
1183 -------
1184 pool_info : `dict` [`str`, `dict` [`str`, Any]]
1185 Mapping between HTCondor slot names and slot information (classAds).
1186 """
1187 if coll is None:
1188 coll = htcondor.Collector()
1189 try:
1190 pool_ads = coll.query(constraint=constraint)
1191 except OSError as ex:
1192 raise RuntimeError(f"Problem querying the Collector. (Constraint='{constraint}')") from ex
1194 pool_info = {}
1195 for slot in pool_ads:
1196 pool_info[slot["name"]] = dict(slot)
1197 _LOG.debug("condor_status returned %d ads", len(pool_info))
1198 return pool_info
1201def update_job_info(job_info, other_info):
1202 """Update results of a job query with results from another query.
1204 Parameters
1205 ----------
1206 job_info : `dict` [`str`, `dict` [`str`, Any]]
1207 Results of the job query that needs to be updated.
1208 other_info : `dict` [`str`, `dict` [`str`, Any]]
1209 Results of the other job query.
1211 Returns
1212 -------
1213 job_info : `dict` [`str`, `dict` [`str`, Any]]
1214 The updated results.
1215 """
1216 for schedd_name, others in other_info.items():
1217 try:
1218 jobs = job_info[schedd_name]
1219 except KeyError:
1220 job_info[schedd_name] = others
1221 else:
1222 for id_, ad in others.items():
1223 jobs.setdefault(id_, {}).update(ad)
1224 return job_info
1227def summary_from_dag(dir_name):
1228 """Build bps_run_summary string from dag file.
1230 Parameters
1231 ----------
1232 dir_name : `str`
1233 Path that includes dag file for a run.
1235 Returns
1236 -------
1237 summary : `str`
1238 Semi-colon separated list of job labels and counts.
1239 (Same format as saved in dag classad).
1240 job_name_to_pipetask : `dict` [`str`, `str`]
1241 Mapping of job names to job labels.
1242 """
1243 dag = next(Path(dir_name).glob("*.dag"))
1245 # Later code depends upon insertion order
1246 counts = defaultdict(int)
1247 job_name_to_pipetask = {}
1248 try:
1249 with open(dag) as fh:
1250 for line in fh:
1251 if line.startswith("JOB"):
1252 m = re.match(r"JOB ([^\s]+) jobs/([^/]+)/", line)
1253 if m:
1254 label = m.group(2)
1255 if label == "init":
1256 label = "pipetaskInit"
1257 job_name_to_pipetask[m.group(1)] = label
1258 counts[label] += 1
1259 else: # Check if Pegasus submission
1260 m = re.match(r"JOB ([^\s]+) ([^\s]+)", line)
1261 if m:
1262 label = pegasus_name_to_label(m.group(1))
1263 job_name_to_pipetask[m.group(1)] = label
1264 counts[label] += 1
1265 else:
1266 _LOG.warning("Parse DAG: unmatched job line: %s", line)
1267 elif line.startswith("FINAL"):
1268 m = re.match(r"FINAL ([^\s]+) jobs/([^/]+)/", line)
1269 if m:
1270 label = m.group(2)
1271 job_name_to_pipetask[m.group(1)] = label
1272 counts[label] += 1
1274 except (OSError, PermissionError, StopIteration):
1275 pass
1277 summary = ";".join([f"{name}:{counts[name]}" for name in counts])
1278 _LOG.debug("summary_from_dag: %s %s", summary, job_name_to_pipetask)
1279 return summary, job_name_to_pipetask
1282def pegasus_name_to_label(name):
1283 """Convert pegasus job name to a label for the report.
1285 Parameters
1286 ----------
1287 name : `str`
1288 Name of job.
1290 Returns
1291 -------
1292 label : `str`
1293 Label for job.
1294 """
1295 label = "UNK"
1296 if name.startswith("create_dir") or name.startswith("stage_in") or name.startswith("stage_out"):
1297 label = "pegasus"
1298 else:
1299 m = re.match(r"pipetask_(\d+_)?([^_]+)", name)
1300 if m:
1301 label = m.group(2)
1302 if label == "init":
1303 label = "pipetaskInit"
1305 return label
1308def read_dag_status(wms_path):
1309 """Read the node status file for DAG summary information.
1311 Parameters
1312 ----------
1313 wms_path : `str`
1314 Path that includes node status file for a run.
1316 Returns
1317 -------
1318 dag_ad : `dict` [`str`, Any]
1319 DAG summary information.
1320 """
1321 dag_ad = {}
1323 # While this is probably more up to date than dag classad, only read from
1324 # file if need to.
1325 try:
1326 try:
1327 node_stat_file = next(Path(wms_path).glob("*.node_status"))
1328 _LOG.debug("Reading Node Status File %s", node_stat_file)
1329 with open(node_stat_file) as infh:
1330 dag_ad = classad.parseNext(infh) # pylint: disable=E1101
1331 except StopIteration:
1332 pass
1334 if not dag_ad:
1335 # Pegasus check here
1336 try:
1337 metrics_file = next(Path(wms_path).glob("*.dag.metrics"))
1338 with open(metrics_file) as infh:
1339 metrics = json.load(infh)
1340 dag_ad["NodesTotal"] = metrics.get("jobs", 0)
1341 dag_ad["NodesFailed"] = metrics.get("jobs_failed", 0)
1342 dag_ad["NodesDone"] = metrics.get("jobs_succeeded", 0)
1343 dag_ad["pegasus_version"] = metrics.get("planner_version", "")
1344 except StopIteration:
1345 try:
1346 metrics_file = next(Path(wms_path).glob("*.metrics"))
1347 with open(metrics_file) as infh:
1348 metrics = json.load(infh)
1349 dag_ad["NodesTotal"] = metrics["wf_metrics"]["total_jobs"]
1350 dag_ad["pegasus_version"] = metrics.get("version", "")
1351 except StopIteration:
1352 pass
1353 except (OSError, PermissionError):
1354 pass
1356 _LOG.debug("read_dag_status: %s", dag_ad)
1357 return dict(dag_ad)
1360def read_node_status(wms_path):
1361 """Read entire node status file.
1363 Parameters
1364 ----------
1365 wms_path : `str`
1366 Path that includes node status file for a run.
1368 Returns
1369 -------
1370 jobs : `dict` [`str`, Any]
1371 DAG summary information compiled from the node status file combined
1372 with the information found in the node event log.
1374 Currently, if the same job attribute is found in both files, its value
1375 from the event log takes precedence over the value from the node status
1376 file.
1377 """
1378 # Get jobid info from other places to fill in gaps in info from node_status
1379 _, job_name_to_pipetask = summary_from_dag(wms_path)
1380 wms_workflow_id, loginfo = read_dag_log(wms_path)
1381 loginfo = read_dag_nodes_log(wms_path)
1382 _LOG.debug("loginfo = %s", loginfo)
1383 job_name_to_id = {}
1384 for jid, jinfo in loginfo.items():
1385 if "LogNotes" in jinfo:
1386 m = re.match(r"DAG Node: ([^\s]+)", jinfo["LogNotes"])
1387 if m:
1388 job_name_to_id[m.group(1)] = jid
1389 jinfo["DAGNodeName"] = m.group(1)
1391 try:
1392 node_status = next(Path(wms_path).glob("*.node_status"))
1393 except StopIteration:
1394 return loginfo
1396 jobs = {}
1397 fake_id = -1.0 # For nodes that do not yet have a job id, give fake one
1398 try:
1399 with open(node_status) as fh:
1400 ads = classad.parseAds(fh)
1402 for jclassad in ads:
1403 if jclassad["Type"] == "DagStatus":
1404 # skip DAG summary
1405 pass
1406 elif "Node" not in jclassad:
1407 if jclassad["Type"] != "StatusEnd":
1408 _LOG.debug("Key 'Node' not in classad: %s", jclassad)
1409 break
1410 else:
1411 if jclassad["Node"] in job_name_to_pipetask:
1412 try:
1413 label = job_name_to_pipetask[jclassad["Node"]]
1414 except KeyError:
1415 _LOG.error("%s not in %s", jclassad["Node"], job_name_to_pipetask.keys())
1416 raise
1417 elif "_" in jclassad["Node"]:
1418 label = jclassad["Node"].split("_")[1]
1419 else:
1420 label = jclassad["Node"]
1422 # Make job info as if came from condor_q
1423 if jclassad["Node"] in job_name_to_id:
1424 job_id = str(job_name_to_id[jclassad["Node"]])
1425 else:
1426 job_id = str(fake_id)
1427 fake_id -= 1
1429 job = dict(jclassad)
1430 job["ClusterId"] = int(float(job_id))
1431 job["DAGManJobID"] = wms_workflow_id
1432 job["DAGNodeName"] = jclassad["Node"]
1433 job["bps_job_label"] = label
1435 jobs[job_id] = job
1436 try:
1437 jobs[job_id] |= loginfo[job_id]
1438 except KeyError:
1439 pass
1440 except (OSError, PermissionError):
1441 pass
1443 return jobs
1446def read_dag_log(wms_path):
1447 """Read job information from the DAGMan log file.
1449 Parameters
1450 ----------
1451 wms_path : `str`
1452 Path containing the DAGMan log file.
1454 Returns
1455 -------
1456 wms_workflow_id : `str`
1457 HTCondor job id (i.e., <ClusterId>.<ProcId>) of the DAGMan job.
1458 dag_info : `dict` [`str`, `~collections.abc.Any`]
1459 HTCondor job information read from the log file mapped to HTCondor
1460 job id.
1462 Raises
1463 ------
1464 FileNotFoundError
1465 If cannot find DAGMan log in given wms_path.
1466 """
1467 wms_workflow_id = 0
1468 dag_info = {}
1470 path = Path(wms_path)
1471 if path.exists():
1472 try:
1473 filename = next(path.glob("*.dag.dagman.log"))
1474 except StopIteration as exc:
1475 raise FileNotFoundError(f"DAGMan log not found in {wms_path}") from exc
1476 _LOG.debug("dag node log filename: %s", filename)
1478 info = {}
1479 job_event_log = htcondor.JobEventLog(str(filename))
1480 for event in job_event_log.events(stop_after=0):
1481 id_ = f"{event['Cluster']}.{event['Proc']}"
1482 if id_ not in info:
1483 info[id_] = {}
1484 wms_workflow_id = id_ # taking last job id in case of restarts
1485 info[id_].update(event)
1486 info[id_][f"{event.type.name.lower()}_time"] = event["EventTime"]
1488 # only save latest DAG job
1489 dag_info = {wms_workflow_id: info[wms_workflow_id]}
1490 for job in dag_info.values():
1491 _tweak_log_info(filename, job)
1493 return wms_workflow_id, dag_info
1496def read_dag_nodes_log(wms_path):
1497 """Read job information from the DAGMan nodes log file.
1499 Parameters
1500 ----------
1501 wms_path : `str`
1502 Path containing the DAGMan nodes log file.
1504 Returns
1505 -------
1506 info : `dict` [`str`, Any]
1507 HTCondor job information read from the log file mapped to HTCondor
1508 job id.
1510 Raises
1511 ------
1512 FileNotFoundError
1513 If cannot find DAGMan node log in given wms_path.
1514 """
1515 try:
1516 filename = next(Path(wms_path).glob("*.dag.nodes.log"))
1517 except StopIteration as exc:
1518 raise FileNotFoundError(f"DAGMan node log not found in {wms_path}") from exc
1519 _LOG.debug("dag node log filename: %s", filename)
1521 info = {}
1522 job_event_log = htcondor.JobEventLog(str(filename))
1523 for event in job_event_log.events(stop_after=0):
1524 id_ = f"{event['Cluster']}.{event['Proc']}"
1525 if id_ not in info:
1526 info[id_] = {}
1527 info[id_].update(event)
1528 info[id_][f"{event.type.name.lower()}_time"] = event["EventTime"]
1530 # Add more condor_q-like info to info parsed from log file.
1531 for job in info.values():
1532 _tweak_log_info(filename, job)
1534 return info
1537def read_dag_info(wms_path):
1538 """Read custom DAGMan job information from the file.
1540 Parameters
1541 ----------
1542 wms_path : `str`
1543 Path containing the file with the DAGMan job info.
1545 Returns
1546 -------
1547 dag_info : `dict` [`str`, `dict` [`str`, Any]]
1548 HTCondor job information.
1550 Raises
1551 ------
1552 FileNotFoundError
1553 If cannot find DAGMan job info file in the given location.
1554 """
1555 try:
1556 filename = next(Path(wms_path).glob("*.info.json"))
1557 except StopIteration as exc:
1558 raise FileNotFoundError(f"File with DAGMan job information not found in {wms_path}") from exc
1559 _LOG.debug("DAGMan job information filename: %s", filename)
1560 try:
1561 with open(filename) as fh:
1562 dag_info = json.load(fh)
1563 except (OSError, PermissionError) as exc:
1564 _LOG.debug("Retrieving DAGMan job information failed: %s", exc)
1565 dag_info = {}
1566 return dag_info
1569def write_dag_info(filename, dag_info):
1570 """Write custom job information about DAGMan job.
1572 Parameters
1573 ----------
1574 filename : `str`
1575 Name of the file where the information will be stored.
1576 dag_info : `dict` [`str` `dict` [`str`, Any]]
1577 Information about the DAGMan job.
1578 """
1579 schedd_name = next(iter(dag_info))
1580 dag_id = next(iter(dag_info[schedd_name]))
1581 dag_ad = dag_info[schedd_name][dag_id]
1582 try:
1583 with open(filename, "w") as fh:
1584 info = {
1585 schedd_name: {
1586 dag_id: {"ClusterId": dag_ad["ClusterId"], "GlobalJobId": dag_ad["GlobalJobId"]}
1587 }
1588 }
1589 json.dump(info, fh)
1590 except (KeyError, OSError, PermissionError) as exc:
1591 _LOG.debug("Persisting DAGMan job information failed: %s", exc)
1594def _tweak_log_info(filename, job):
1595 """Massage the given job info has same structure as if came from condor_q.
1597 Parameters
1598 ----------
1599 filename : `pathlib.Path`
1600 Name of the DAGMan log.
1601 job : `dict` [ `str`, Any ]
1602 A mapping between HTCondor job id and job information read from
1603 the log.
1604 """
1605 _LOG.debug("_tweak_log_info: %s %s", filename, job)
1607 try:
1608 job["ClusterId"] = job["Cluster"]
1609 job["ProcId"] = job["Proc"]
1610 job["Iwd"] = str(filename.parent)
1611 job["Owner"] = filename.owner()
1613 match job["MyType"]:
1614 case "ExecuteEvent":
1615 job["JobStatus"] = JobStatus.RUNNING
1616 case "JobTerminatedEvent" | "PostScriptTerminatedEvent":
1617 job["JobStatus"] = JobStatus.COMPLETED
1618 case "SubmitEvent":
1619 job["JobStatus"] = JobStatus.IDLE
1620 case "JobAbortedEvent":
1621 job["JobStatus"] = JobStatus.REMOVED
1622 case "JobHeldEvent":
1623 job["JobStatus"] = JobStatus.HELD
1624 case _:
1625 _LOG.debug("Unknown log event type: %s", job["MyType"])
1626 job["JobStatus"] = JobStatus.UNEXPANDED
1628 if job["JobStatus"] in {JobStatus.COMPLETED, JobStatus.HELD}:
1629 new_job = HTC_JOB_AD_HANDLERS.handle(job)
1630 if new_job is not None:
1631 job = new_job
1632 else:
1633 _LOG.error("Could not determine exit status for job '%s.%s'", job["ClusterId"], job["ProcId"])
1635 except KeyError as e:
1636 _LOG.error("Missing key %s in job: %s", str(e), job)
1637 raise
1640def htc_check_dagman_output(wms_path):
1641 """Check the DAGMan output for error messages.
1643 Parameters
1644 ----------
1645 wms_path : `str`
1646 Directory containing the DAGman output file.
1648 Returns
1649 -------
1650 message : `str`
1651 Message containing error messages from the DAGMan output. Empty
1652 string if no messages.
1654 Raises
1655 ------
1656 FileNotFoundError
1657 If cannot find DAGMan standard output file in given wms_path.
1658 """
1659 try:
1660 filename = next(Path(wms_path).glob("*.dag.dagman.out"))
1661 except StopIteration as exc:
1662 raise FileNotFoundError(f"DAGMan standard output file not found in {wms_path}") from exc
1663 _LOG.debug("dag output filename: %s", filename)
1665 message = ""
1666 try:
1667 with open(filename) as fh:
1668 last_submit_failed = ""
1669 for line in fh:
1670 m = re.match(r"(\d\d/\d\d/\d\d \d\d:\d\d:\d\d) Job submit try \d+/\d+ failed", line)
1671 if m:
1672 last_submit_failed = m.group(1)
1673 if last_submit_failed:
1674 message = f"Warn: Job submission issues (last: {last_submit_failed})"
1675 except (OSError, PermissionError):
1676 message = f"Warn: Could not read dagman output file from {wms_path}."
1677 return message