-
Notifications
You must be signed in to change notification settings - Fork 353
/
scheduler_models.py
1317 lines (1013 loc) · 46.9 KB
/
scheduler_models.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
"""Database model classes for the scheduler.
Contains model classes abstracting the various DB tables used by the scheduler.
These overlap the Django models in basic functionality, but were written before
the Django models existed and have not yet been phased out. Some of them
(particularly HostQueueEntry and Job) have considerable scheduler-specific logic
which would probably be ill-suited for inclusion in the general Django model
classes.
Globals:
_notify_email_statuses: list of HQE statuses. each time a single HQE reaches
one of these statuses, an email will be sent to the job's email_list.
comes from global_config.
_base_url: URL to the local AFE server, used to construct URLs for emails.
_db: DatabaseConnection for this module.
_drone_manager: reference to global DroneManager instance.
"""
import datetime, itertools, logging, os, re, sys, time, weakref
from django.db import connection
from autotest.client.shared import global_config, host_protections, utils
from autotest.frontend.afe import models, model_attributes
from autotest.database import database_connection
from autotest.scheduler import drone_manager, email_manager
from autotest.scheduler import scheduler_config
_notify_email_statuses = []
_base_url = None
_db = None
_drone_manager = None
def initialize():
global _db
_db = database_connection.DatabaseConnection('AUTOTEST_WEB')
_db.connect(db_type='django')
notify_statuses_list = global_config.global_config.get_config_value(
scheduler_config.CONFIG_SECTION, "notify_email_statuses",
default='')
global _notify_email_statuses
_notify_email_statuses = [status for status in
re.split(r'[\s,;:]', notify_statuses_list.lower())
if status]
# AUTOTEST_WEB.base_url is still a supported config option as some people
# may wish to override the entire url.
global _base_url
config_base_url = global_config.global_config.get_config_value(
scheduler_config.CONFIG_SECTION, 'base_url', default='')
if config_base_url:
_base_url = config_base_url
else:
# For the common case of everything running on a single server you
# can just set the hostname in a single place in the config file.
server_name = global_config.global_config.get_config_value(
'SERVER', 'hostname')
if not server_name:
logging.critical('[SERVER] hostname missing from the config file.')
sys.exit(1)
_base_url = 'http://%s/afe/' % server_name
initialize_globals()
def initialize_globals():
global _drone_manager
_drone_manager = drone_manager.instance()
class DelayedCallTask(object):
"""
A task object like AgentTask for an Agent to run that waits for the
specified amount of time to have elapsed before calling the supplied
callback once and finishing. If the callback returns anything, it is
assumed to be a new Agent instance and will be added to the dispatcher.
@attribute end_time: The absolute posix time after which this task will
call its callback when it is polled and be finished.
Also has all attributes required by the Agent class.
"""
def __init__(self, delay_seconds, callback, now_func=None):
"""
@param delay_seconds: The delay in seconds from now that this task
will call the supplied callback and be done.
@param callback: A callable to be called by this task once after at
least delay_seconds time has elapsed. It must return None
or a new Agent instance.
@param now_func: A time.time like function. Default: time.time.
Used for testing.
"""
assert delay_seconds > 0
assert callable(callback)
if not now_func:
now_func = time.time
self._now_func = now_func
self._callback = callback
self.end_time = self._now_func() + delay_seconds
# These attributes are required by Agent.
self.aborted = False
self.host_ids = ()
self.success = False
self.queue_entry_ids = ()
self.num_processes = 0
def poll(self):
if not self.is_done() and self._now_func() >= self.end_time:
self._callback()
self.success = True
def is_done(self):
return self.success or self.aborted
def abort(self):
self.aborted = True
class DBError(Exception):
"""Raised by the DBObject constructor when its select fails."""
class DBObject(object):
"""A miniature object relational model for the database."""
# Subclasses MUST override these:
_table_name = ''
_fields = ()
# A mapping from (type, id) to the instance of the object for that
# particular id. This prevents us from creating new Job() and Host()
# instances for every HostQueueEntry object that we instantiate as
# multiple HQEs often share the same Job.
_instances_by_type_and_id = weakref.WeakValueDictionary()
_initialized = False
def __new__(cls, id=None, **kwargs):
"""
Look to see if we already have an instance for this particular type
and id. If so, use it instead of creating a duplicate instance.
"""
if id is not None:
instance = cls._instances_by_type_and_id.get((cls, id))
if instance:
return instance
return super(DBObject, cls).__new__(cls, id=id, **kwargs)
def __init__(self, id=None, row=None, new_record=False, always_query=True):
assert bool(id) or bool(row)
if id is not None and row is not None:
assert id == row[0]
assert self._table_name, '_table_name must be defined in your class'
assert self._fields, '_fields must be defined in your class'
if not new_record:
if self._initialized and not always_query:
return # We've already been initialized.
if id is None:
id = row[0]
# Tell future constructors to use us instead of re-querying while
# this instance is still around.
self._instances_by_type_and_id[(type(self), id)] = self
self.__table = self._table_name
self.__new_record = new_record
if row is None:
row = self._fetch_row_from_db(id)
if self._initialized:
differences = self._compare_fields_in_row(row)
if differences:
logging.warn(
'initialized %s %s instance requery is updating: %s',
type(self), self.id, differences)
self._update_fields_from_row(row)
self._initialized = True
@classmethod
def _clear_instance_cache(cls):
"""Used for testing, clear the internal instance cache."""
cls._instances_by_type_and_id.clear()
def _fetch_row_from_db(self, row_id):
sql = 'SELECT * FROM %s WHERE ID=%%s' % self.__table
rows = _db.execute(sql, (row_id,))
if not rows:
raise DBError("row not found (table=%s, row id=%s)"
% (self.__table, row_id))
return rows[0]
def _assert_row_length(self, row):
assert len(row) == len(self._fields), (
"table = %s, row = %s/%d, fields = %s/%d" % (
self.__table, row, len(row), self._fields, len(self._fields)))
def _compare_fields_in_row(self, row):
"""
Given a row as returned by a SELECT query, compare it to our existing in
memory fields. Fractional seconds are stripped from datetime values
before comparison.
@param row - A sequence of values corresponding to fields named in
The class attribute _fields.
@returns A dictionary listing the differences keyed by field name
containing tuples of (current_value, row_value).
"""
self._assert_row_length(row)
differences = {}
datetime_cmp_fmt = '%Y-%m-%d %H:%M:%S' # Leave off the microseconds.
for field, row_value in itertools.izip(self._fields, row):
current_value = getattr(self, field)
if (isinstance(current_value, datetime.datetime)
and isinstance(row_value, datetime.datetime)):
current_value = current_value.strftime(datetime_cmp_fmt)
row_value = row_value.strftime(datetime_cmp_fmt)
if current_value != row_value:
differences[field] = (current_value, row_value)
return differences
def _update_fields_from_row(self, row):
"""
Update our field attributes using a single row returned by SELECT.
@param row - A sequence of values corresponding to fields named in
the class fields list.
"""
self._assert_row_length(row)
self._valid_fields = set()
for field, value in itertools.izip(self._fields, row):
setattr(self, field, value)
self._valid_fields.add(field)
self._valid_fields.remove('id')
def update_from_database(self):
assert self.id is not None
row = self._fetch_row_from_db(self.id)
self._update_fields_from_row(row)
def count(self, where, table = None):
if not table:
table = self.__table
rows = _db.execute("""
SELECT count(*) FROM %s
WHERE %s
""" % (table, where))
assert len(rows) == 1
return int(rows[0][0])
def update_field(self, field, value):
assert field in self._valid_fields
if getattr(self, field) == value:
return
query = "UPDATE %s SET %s = %%s WHERE id = %%s" % (self.__table, field)
_db.execute(query, (value, self.id))
setattr(self, field, value)
def save(self):
if self.__new_record:
keys = self._fields[1:] # avoid id
columns = ','.join([str(key) for key in keys])
values = []
for key in keys:
value = getattr(self, key)
if value is None:
values.append('NULL')
else:
values.append('"%s"' % value)
values_str = ','.join(values)
query = ('INSERT INTO %s (%s) VALUES (%s)' %
(self.__table, columns, values_str))
_db.execute(query)
# Update our id to the one the database just assigned to us.
self.id = _db.execute('SELECT LAST_INSERT_ID()')[0][0]
def delete(self):
self._instances_by_type_and_id.pop((type(self), id), None)
self._initialized = False
self._valid_fields.clear()
query = 'DELETE FROM %s WHERE id=%%s' % self.__table
_db.execute(query, (self.id,))
@staticmethod
def _prefix_with(string, prefix):
if string:
string = prefix + string
return string
@classmethod
def fetch(cls, where='', params=(), joins='', order_by=''):
"""
Construct instances of our class based on the given database query.
@yields One class instance for each row fetched.
"""
order_by = cls._prefix_with(order_by, 'ORDER BY ')
where = cls._prefix_with(where, 'WHERE ')
query = ('SELECT %(table)s.* FROM %(table)s %(joins)s '
'%(where)s %(order_by)s' % {'table' : cls._table_name,
'joins' : joins,
'where' : where,
'order_by' : order_by})
rows = _db.execute(query, params)
return [cls(id=row[0], row=row) for row in rows]
class IneligibleHostQueue(DBObject):
_table_name = 'afe_ineligible_host_queues'
_fields = ('id', 'job_id', 'host_id')
class AtomicGroup(DBObject):
_table_name = 'afe_atomic_groups'
_fields = ('id', 'name', 'description', 'max_number_of_machines',
'invalid')
class Label(DBObject):
_table_name = 'afe_labels'
_fields = ('id', 'name', 'kernel_config', 'platform', 'invalid',
'only_if_needed', 'atomic_group_id')
def __repr__(self):
return 'Label(name=%r, id=%d, atomic_group_id=%r)' % (
self.name, self.id, self.atomic_group_id)
class Host(DBObject):
_table_name = 'afe_hosts'
_fields = ('id', 'hostname', 'locked', 'synch_id', 'status',
'invalid', 'protection', 'locked_by_id', 'lock_time', 'dirty')
def set_status(self,status):
logging.info('%s -> %s', self.hostname, status)
self.update_field('status',status)
def platform_and_labels(self):
"""
Returns a tuple (platform_name, list_of_all_label_names).
"""
rows = _db.execute("""
SELECT afe_labels.name, afe_labels.platform
FROM afe_labels
INNER JOIN afe_hosts_labels ON
afe_labels.id = afe_hosts_labels.label_id
WHERE afe_hosts_labels.host_id = %s
ORDER BY afe_labels.name
""", (self.id,))
platform = None
all_labels = []
for label_name, is_platform in rows:
if is_platform:
platform = label_name
all_labels.append(label_name)
return platform, all_labels
_ALPHANUM_HOST_RE = re.compile(r'^([a-z-]+)(\d+)$', re.IGNORECASE)
@classmethod
def cmp_for_sort(cls, a, b):
"""
A comparison function for sorting Host objects by hostname.
This strips any trailing numeric digits, ignores leading 0s and
compares hostnames by the leading name and the trailing digits as a
number. If both hostnames do not match this pattern, they are simply
compared as lower case strings.
Example of how hostnames will be sorted:
alice, host1, host2, host09, host010, host10, host11, yolkfolk
This hopefully satisfy most people's hostname sorting needs regardless
of their exact naming schemes. Nobody sane should have both a host10
and host010 (but the algorithm works regardless).
"""
lower_a = a.hostname.lower()
lower_b = b.hostname.lower()
match_a = cls._ALPHANUM_HOST_RE.match(lower_a)
match_b = cls._ALPHANUM_HOST_RE.match(lower_b)
if match_a and match_b:
name_a, number_a_str = match_a.groups()
name_b, number_b_str = match_b.groups()
number_a = int(number_a_str.lstrip('0'))
number_b = int(number_b_str.lstrip('0'))
result = cmp((name_a, number_a), (name_b, number_b))
if result == 0 and lower_a != lower_b:
# If they compared equal above but the lower case names are
# indeed different, don't report equality. abc012 != abc12.
return cmp(lower_a, lower_b)
return result
else:
return cmp(lower_a, lower_b)
class HostQueueEntry(DBObject):
_table_name = 'afe_host_queue_entries'
_fields = ('id', 'job_id', 'host_id', 'profile', 'status', 'meta_host',
'active', 'complete', 'deleted', 'execution_subdir',
'atomic_group_id', 'aborted', 'started_on')
def __init__(self, id=None, row=None, **kwargs):
assert id or row
super(HostQueueEntry, self).__init__(id=id, row=row, **kwargs)
self.job = Job(self.job_id)
if self.host_id:
self.host = Host(self.host_id)
else:
self.host = None
if self.atomic_group_id:
self.atomic_group = AtomicGroup(self.atomic_group_id,
always_query=False)
else:
self.atomic_group = None
self.queue_log_path = os.path.join(self.job.tag(),
'queue.log.' + str(self.id))
@classmethod
def clone(cls, template):
"""
Creates a new row using the values from a template instance.
The new instance will not exist in the database or have a valid
id attribute until its save() method is called.
"""
assert isinstance(template, cls)
new_row = [getattr(template, field) for field in cls._fields]
clone = cls(row=new_row, new_record=True)
clone.id = None
return clone
def _view_job_url(self):
return "%s#tab_id=view_job&object_id=%s" % (_base_url, self.job.id)
def get_labels(self):
"""
Get all labels associated with this host queue entry (either via the
meta_host or as a job dependency label). The labels yielded are not
guaranteed to be unique.
@yields Label instances associated with this host_queue_entry.
"""
if self.meta_host:
yield Label(id=self.meta_host, always_query=False)
labels = Label.fetch(
joins="JOIN afe_jobs_dependency_labels AS deps "
"ON (afe_labels.id = deps.label_id)",
where="deps.job_id = %d" % self.job.id)
for label in labels:
yield label
def set_host(self, host):
if host:
logging.info('Assigning host %s to entry %s', host.hostname, self)
self.queue_log_record('Assigning host ' + host.hostname)
self.update_field('host_id', host.id)
self.block_host(host.id)
else:
logging.info('Releasing host from %s', self)
self.queue_log_record('Releasing host')
self.unblock_host(self.host.id)
self.update_field('host_id', None)
self.host = host
def queue_log_record(self, log_line):
now = str(datetime.datetime.now())
_drone_manager.write_lines_to_file(self.queue_log_path,
[now + ' ' + log_line])
def block_host(self, host_id):
logging.info("creating block %s/%s", self.job.id, host_id)
row = [0, self.job.id, host_id]
block = IneligibleHostQueue(row=row, new_record=True)
block.save()
def unblock_host(self, host_id):
logging.info("removing block %s/%s", self.job.id, host_id)
blocks = IneligibleHostQueue.fetch(
'job_id=%d and host_id=%d' % (self.job.id, host_id))
for block in blocks:
block.delete()
def set_execution_subdir(self, subdir=None):
if subdir is None:
assert self.host
subdir = self.host.hostname
self.update_field('execution_subdir', subdir)
def _get_hostname(self):
if self.host:
return self.host.hostname
return 'no host'
def __str__(self):
flags = []
if self.active:
flags.append('active')
if self.complete:
flags.append('complete')
if self.deleted:
flags.append('deleted')
if self.aborted:
flags.append('aborted')
flags_str = ','.join(flags)
if flags_str:
flags_str = ' [%s]' % flags_str
return "%s/%d (%d) %s%s" % (self._get_hostname(), self.job.id, self.id,
self.status, flags_str)
def set_status(self, status):
logging.info("%s -> %s", self, status)
self.update_field('status', status)
active = (status in models.HostQueueEntry.ACTIVE_STATUSES)
complete = (status in models.HostQueueEntry.COMPLETE_STATUSES)
assert not (active and complete)
self.update_field('active', active)
self.update_field('complete', complete)
if complete:
self._on_complete(status)
self._email_on_job_complete()
should_email_status = (status.lower() in _notify_email_statuses or
'all' in _notify_email_statuses)
if should_email_status:
self._email_on_status(status)
def _on_complete(self, status):
if status is not models.HostQueueEntry.Status.ABORTED:
self.job.stop_if_necessary()
if not self.execution_subdir:
return
# unregister any possible pidfiles associated with this queue entry
for pidfile_name in drone_manager.ALL_PIDFILE_NAMES:
pidfile_id = _drone_manager.get_pidfile_id_from(
self.execution_path(), pidfile_name=pidfile_name)
_drone_manager.unregister_pidfile(pidfile_id)
def _get_status_email_contents(self, status, summary=None, hostname=None):
"""
Gather info for the status notification e-mails.
If needed, we could start using the Django templating engine to create
the subject and the e-mail body, but that doesn't seem necessary right
now.
@param status: Job status text. Mandatory.
@param summary: Job summary text. Optional.
@param hostname: A hostname for the job. Optional.
@return: Tuple (subject, body) for the notification e-mail.
"""
job_stats = Job(id=self.job.id).get_execution_details()
subject = ('Autotest | Job ID: %s "%s" | Status: %s ' %
(self.job.id, self.job.name, status))
if hostname is not None:
subject += '| Hostname: %s ' % hostname
if status not in ["1 Failed", "Failed"]:
subject += '| Success Rate: %.2f %%' % job_stats['success_rate']
body = "Job ID: %s\n" % self.job.id
body += "Job name: %s\n" % self.job.name
if hostname is not None:
body += "Host: %s\n" % hostname
if summary is not None:
body += "Summary: %s\n" % summary
body += "Status: %s\n" % status
body += "Results interface URL: %s\n" % self._view_job_url()
body += "Execution time (HH:MM:SS): %s\n" % job_stats['execution_time']
if int(job_stats['total_executed']) > 0:
body += "User tests executed: %s\n" % job_stats['total_executed']
body += "User tests passed: %s\n" % job_stats['total_passed']
body += "User tests failed: %s\n" % job_stats['total_failed']
body += ("User tests success rate: %.2f %%\n" %
job_stats['success_rate'])
if job_stats['failed_rows']:
body += "Failures:\n"
body += job_stats['failed_rows']
return subject, body
def _email_on_status(self, status):
hostname = self._get_hostname()
subject, body = self._get_status_email_contents(status, None, hostname)
email_manager.manager.send_email(self.job.email_list, subject, body)
def _email_on_job_complete(self):
if not self.job.is_finished():
return
summary = []
hosts_queue = HostQueueEntry.fetch('job_id = %s' % self.job.id)
for queue_entry in hosts_queue:
summary.append("Host: %s Status: %s" %
(queue_entry._get_hostname(),
queue_entry.status))
summary = "\n".join(summary)
status_counts = models.Job.objects.get_status_counts(
[self.job.id])[self.job.id]
status = ', '.join('%d %s' % (count, status) for status, count
in status_counts.iteritems())
subject, body = self._get_status_email_contents(status, summary, None)
email_manager.manager.send_email(self.job.email_list, subject, body)
def schedule_pre_job_tasks(self):
logging.info("%s/%s/%s (job %s, entry %s) scheduled on %s, status=%s",
self.job.name, self.meta_host, self.atomic_group_id,
self.job.id, self.id, self.host.hostname, self.status)
self._do_schedule_pre_job_tasks()
def _do_schedule_pre_job_tasks(self):
# Every host goes thru the Verifying stage (which may or may not
# actually do anything as determined by get_pre_job_tasks).
self.set_status(models.HostQueueEntry.Status.VERIFYING)
self.job.schedule_pre_job_tasks(queue_entry=self)
def requeue(self):
assert self.host
self.set_status(models.HostQueueEntry.Status.QUEUED)
self.update_field('started_on', None)
# verify/cleanup failure sets the execution subdir, so reset it here
self.set_execution_subdir('')
if self.meta_host:
self.set_host(None)
@property
def aborted_by(self):
self._load_abort_info()
return self._aborted_by
@property
def aborted_on(self):
self._load_abort_info()
return self._aborted_on
def _load_abort_info(self):
""" Fetch info about who aborted the job. """
if hasattr(self, "_aborted_by"):
return
rows = _db.execute("""
SELECT afe_users.login,
afe_aborted_host_queue_entries.aborted_on
FROM afe_aborted_host_queue_entries
INNER JOIN afe_users
ON afe_users.id = afe_aborted_host_queue_entries.aborted_by_id
WHERE afe_aborted_host_queue_entries.queue_entry_id = %s
""", (self.id,))
if rows:
self._aborted_by, self._aborted_on = rows[0]
else:
self._aborted_by = self._aborted_on = None
def on_pending(self):
"""
Called when an entry in a synchronous job has passed verify. If the
job is ready to run, sets the entries to STARTING. Otherwise, it leaves
them in PENDING.
"""
self.set_status(models.HostQueueEntry.Status.PENDING)
self.host.set_status(models.Host.Status.PENDING)
# Some debug code here: sends an email if an asynchronous job does not
# immediately enter Starting.
# TODO: Remove this once we figure out why asynchronous jobs are getting
# stuck in Pending.
self.job.run_if_ready(queue_entry=self)
if (self.job.synch_count == 1 and
self.status == models.HostQueueEntry.Status.PENDING):
subject = 'Job %s (id %s)' % (self.job.name, self.job.id)
message = 'Asynchronous job stuck in Pending'
email_manager.manager.enqueue_notify_email(subject, message)
def abort(self, dispatcher):
assert self.aborted and not self.complete
Status = models.HostQueueEntry.Status
if self.status in (Status.GATHERING, Status.PARSING, Status.ARCHIVING):
# do nothing; post-job tasks will finish and then mark this entry
# with status "Aborted" and take care of the host
return
if self.status in (Status.STARTING, Status.PENDING, Status.RUNNING,
Status.WAITING):
assert not dispatcher.get_agents_for_entry(self)
self.host.set_status(models.Host.Status.READY)
elif self.status == Status.VERIFYING:
models.SpecialTask.objects.create(
task=models.SpecialTask.Task.CLEANUP,
host=models.Host.objects.get(id=self.host.id),
requested_by=self.job.owner_model())
self.set_status(Status.ABORTED)
self.job.abort_delay_ready_task()
def get_group_name(self):
atomic_group = self.atomic_group
if not atomic_group:
return ''
# Look at any meta_host and dependency labels and pick the first
# one that also specifies this atomic group. Use that label name
# as the group name if possible (it is more specific).
for label in self.get_labels():
if label.atomic_group_id:
assert label.atomic_group_id == atomic_group.id
return label.name
return atomic_group.name
def execution_tag(self):
assert self.execution_subdir
return "%s/%s" % (self.job.tag(), self.execution_subdir)
def execution_path(self):
return self.execution_tag()
def set_started_on_now(self):
self.update_field('started_on', datetime.datetime.now())
def is_hostless(self):
return (self.host_id is None
and self.meta_host is None
and self.atomic_group_id is None)
class Job(DBObject):
_table_name = 'afe_jobs'
_fields = ('id', 'owner', 'name', 'priority', 'control_file',
'control_type', 'created_on', 'synch_count', 'timeout',
'run_verify', 'email_list', 'reboot_before', 'reboot_after',
'parse_failed_repair', 'max_runtime_hrs', 'drone_set_id',
'parameterized_job_id')
# This does not need to be a column in the DB. The delays are likely to
# be configured short. If the scheduler is stopped and restarted in
# the middle of a job's delay cycle, the delay cycle will either be
# repeated or skipped depending on the number of Pending machines found
# when the restarted scheduler recovers to track it. Not a problem.
#
# A reference to the DelayedCallTask that will wake up the job should
# no other HQEs change state in time. Its end_time attribute is used
# by our run_with_ready_delay() method to determine if the wait is over.
_delay_ready_task = None
# TODO(gps): On scheduler start/recovery we need to call HQE.on_pending() on
# all status='Pending' atomic group HQEs incase a delay was running when the
# scheduler was restarted and no more hosts ever successfully exit Verify.
def __init__(self, id=None, row=None, **kwargs):
assert id or row
super(Job, self).__init__(id=id, row=row, **kwargs)
self._owner_model = None # caches model instance of owner
def model(self):
return models.Job.objects.get(id=self.id)
def owner_model(self):
# work around the fact that the Job owner field is a string, not a
# foreign key
if not self._owner_model:
self._owner_model = models.User.objects.get(login=self.owner)
return self._owner_model
def is_server_job(self):
return self.control_type != 2
def tag(self):
return "%s-%s" % (self.id, self.owner)
def get_host_queue_entries(self):
rows = _db.execute("""
SELECT * FROM afe_host_queue_entries
WHERE job_id= %s
""", (self.id,))
entries = [HostQueueEntry(row=i) for i in rows]
assert len(entries)>0
return entries
def get_execution_details(self):
"""
Get test execution details for this job.
@return: Dictionary with test execution details
"""
def _find_test_jobs(rows):
"""
Here we are looking for tests such as SERVER_JOB and CLIENT_JOB.*
Those are autotest 'internal job' tests, so they should not be
counted when evaluating the test stats.
@param rows: List of rows (matrix) with database results.
"""
job_test_pattern = re.compile('SERVER|CLIENT\\_JOB\.[\d]')
n_test_jobs = 0
for r in rows:
test_name = r[0]
if job_test_pattern.match(test_name):
n_test_jobs += 1
return n_test_jobs
stats = {}
rows = _db.execute("""
SELECT t.test, s.word, t.reason
FROM tko_tests AS t, tko_jobs AS j, tko_status AS s
WHERE t.job_idx = j.job_idx
AND s.status_idx = t.status
AND j.afe_job_id = %s
ORDER BY t.reason
""" % self.id)
failed_rows = [r for r in rows if not r[1] == 'GOOD']
n_test_jobs = _find_test_jobs(rows)
n_test_jobs_failed = _find_test_jobs(failed_rows)
total_executed = len(rows) - n_test_jobs
total_failed = len(failed_rows) - n_test_jobs_failed
if total_executed > 0:
success_rate = 100 - ((total_failed / float(total_executed)) * 100)
else:
success_rate = 0
stats['total_executed'] = total_executed
stats['total_failed'] = total_failed
stats['total_passed'] = total_executed - total_failed
stats['success_rate'] = success_rate
status_header = ("Test Name", "Status", "Reason")
if failed_rows:
stats['failed_rows'] = utils.matrix_to_string(failed_rows,
status_header)
else:
stats['failed_rows'] = ''
time_row = _db.execute("""
SELECT started_time, finished_time
FROM tko_jobs
WHERE afe_job_id = %s
""" % self.id)
if time_row:
t_begin, t_end = time_row[0]
try:
delta = t_end - t_begin
minutes, seconds = divmod(delta.seconds, 60)
hours, minutes = divmod(minutes, 60)
stats['execution_time'] = ("%02d:%02d:%02d" %
(hours, minutes, seconds))
# One of t_end or t_begin are None
except TypeError:
stats['execution_time'] = '(could not determine)'
else:
stats['execution_time'] = '(none)'
return stats
def set_status(self, status, update_queues=False):
self.update_field('status',status)
if update_queues:
for queue_entry in self.get_host_queue_entries():
queue_entry.set_status(status)
def keyval_dict(self):
return self.model().keyval_dict()
def _atomic_and_has_started(self):
"""
@returns True if any of the HostQueueEntries associated with this job
have entered the Status.STARTING state or beyond.
"""
atomic_entries = models.HostQueueEntry.objects.filter(
job=self.id, atomic_group__isnull=False)
if atomic_entries.count() <= 0:
return False
# These states may *only* be reached if Job.run() has been called.
started_statuses = (models.HostQueueEntry.Status.STARTING,
models.HostQueueEntry.Status.RUNNING,
models.HostQueueEntry.Status.COMPLETED)
started_entries = atomic_entries.filter(status__in=started_statuses)
return started_entries.count() > 0
def _hosts_assigned_count(self):
"""The number of HostQueueEntries assigned a Host for this job."""
entries = models.HostQueueEntry.objects.filter(job=self.id,
host__isnull=False)
return entries.count()
def _pending_count(self):
"""The number of HostQueueEntries for this job in the Pending state."""
pending_entries = models.HostQueueEntry.objects.filter(
job=self.id, status=models.HostQueueEntry.Status.PENDING)
return pending_entries.count()
def _max_hosts_needed_to_run(self, atomic_group):
"""
@param atomic_group: The AtomicGroup associated with this job that we
are using to set an upper bound on the threshold.
@returns The maximum number of HostQueueEntries assigned a Host before
this job can run.
"""
return min(self._hosts_assigned_count(),
atomic_group.max_number_of_machines)