-
Notifications
You must be signed in to change notification settings - Fork 307
/
request.py
2928 lines (2565 loc) · 114 KB
/
request.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
# -*- coding: utf-8 -*-
# Copyright European Organization for Nuclear Research (CERN) since 2012
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import datetime
import itertools
import json
import logging
import math
import random
import threading
import traceback
import uuid
from abc import ABCMeta, abstractmethod
from collections import namedtuple, defaultdict
from collections.abc import Sequence, Mapping, Iterator
from dataclasses import dataclass
from typing import TYPE_CHECKING, Any, Optional, Union
from sqlalchemy import and_, or_, update, select, delete, exists, insert
from sqlalchemy.exc import IntegrityError
from sqlalchemy.orm import aliased
from sqlalchemy.sql.expression import asc, true, false, null, func
from rucio.common.config import config_get_bool, config_get_int
from rucio.common.exception import RequestNotFound, RucioException, UnsupportedOperation, InvalidRSEExpression
from rucio.common.types import InternalAccount, InternalScope
from rucio.common.utils import generate_uuid, chunks
from rucio.core.distance import get_distances
from rucio.core.message import add_message, add_messages
from rucio.core.monitor import MetricManager
from rucio.core.rse import get_rse_attribute, get_rse_name, get_rse_vo, RseData, RseCollection
from rucio.core.rse_expression_parser import parse_expression
from rucio.db.sqla import models, filter_thread_work
from rucio.db.sqla.constants import RequestState, RequestType, LockState, RequestErrMsg, ReplicaState, TransferLimitDirection
from rucio.db.sqla.session import read_session, transactional_session, stream_session
from rucio.db.sqla.util import temp_table_mngr
RequestAndState = namedtuple('RequestAndState', ['request_id', 'request_state'])
if TYPE_CHECKING:
from sqlalchemy.orm import Session
from rucio.rse.protocols.protocol import RSEProtocol
"""
The core request.py is specifically for handling requests.
Requests accessed by external_id (So called transfers), are covered in the core transfer.py
"""
METRICS = MetricManager(module=__name__)
TRANSFER_TIME_BUCKETS = (
10, 30, 60, 5 * 60, 10 * 60, 20 * 60, 40 * 60, 60 * 60, 1.5 * 60 * 60, 3 * 60 * 60, 6 * 60 * 60,
12 * 60 * 60, 24 * 60 * 60, 3 * 24 * 60 * 60, 4 * 24 * 60 * 60, 5 * 24 * 60 * 60,
6 * 24 * 60 * 60, 7 * 24 * 60 * 60, 10 * 24 * 60 * 60, 14 * 24 * 60 * 60, 30 * 24 * 60 * 60,
float('inf')
)
class RequestSource:
def __init__(self, rse: RseData, ranking=None, distance=None, file_path=None, scheme=None, url=None):
self.rse = rse
self.distance = distance if distance is not None else 9999
self.ranking = ranking if ranking is not None else 0
self.file_path = file_path
self.scheme = scheme
self.url = url
def __str__(self):
return "src_rse={}".format(self.rse)
class TransferDestination:
def __init__(self, rse: RseData, scheme):
self.rse = rse
self.scheme = scheme
def __str__(self):
return "dst_rse={}".format(self.rse)
class RequestWithSources:
def __init__(
self,
id_: Optional[str],
request_type: RequestType,
rule_id: Optional[str],
scope: InternalScope,
name: str,
md5: str,
adler32: str,
byte_count: int,
activity: str,
attributes: Optional[Union[str, dict[str, Any]]],
previous_attempt_id: Optional[str],
dest_rse: RseData,
account: InternalAccount,
retry_count: int,
priority: int,
transfertool: str,
requested_at: Optional[datetime.datetime] = None,
):
self.request_id = id_
self.request_type = request_type
self.rule_id = rule_id
self.scope = scope
self.name = name
self.md5 = md5
self.adler32 = adler32
self.byte_count = byte_count
self.activity = activity
self._dict_attributes = None
self._db_attributes = attributes
self.previous_attempt_id = previous_attempt_id
self.dest_rse = dest_rse
self.account = account
self.retry_count = retry_count or 0
self.priority = priority if priority is not None else 3
self.transfertool = transfertool
self.requested_at = requested_at if requested_at else datetime.datetime.utcnow()
self.sources: list[RequestSource] = []
self.requested_source: Optional[RequestSource] = None
def __str__(self):
return "{}({}:{})".format(self.request_id, self.scope, self.name)
@property
def attributes(self):
if self._dict_attributes is None:
self._dict_attributes = self._parse_db_attributes(self._db_attributes)
return self._dict_attributes
@attributes.setter
def attributes(self, db_attributes):
self._dict_attributes = self._parse_db_attributes(db_attributes)
@staticmethod
def _parse_db_attributes(db_attributes):
attr = {}
if db_attributes:
if isinstance(db_attributes, dict):
attr = json.loads(json.dumps(db_attributes))
else:
attr = json.loads(str(db_attributes))
# parse source expression
attr['source_replica_expression'] = attr["source_replica_expression"] if (attr and "source_replica_expression" in attr) else None
attr['allow_tape_source'] = attr["allow_tape_source"] if (attr and "allow_tape_source" in attr) else True
attr['dsn'] = attr["ds_name"] if (attr and "ds_name" in attr) else None
attr['lifetime'] = attr.get('lifetime', -1)
return attr
class DirectTransfer(metaclass=ABCMeta):
"""
The configuration for a direct (non-multi-hop) transfer. It can be a multi-source transfer.
"""
def __init__(self, sources: list[RequestSource], rws: RequestWithSources) -> None:
self.sources: list[RequestSource] = sources
self.rws: RequestWithSources = rws
@property
@abstractmethod
def src(self) -> RequestSource:
pass
@property
@abstractmethod
def dst(self) -> TransferDestination:
pass
@property
@abstractmethod
def dest_url(self) -> str:
pass
@abstractmethod
def source_url(self, source: RequestSource) -> str:
pass
@abstractmethod
def dest_protocol(self) -> "RSEProtocol":
pass
@abstractmethod
def source_protocol(self, source: RequestSource) -> "RSEProtocol":
pass
def should_retry_request(req, retry_protocol_mismatches):
"""
Whether should retry this request.
:param request: Request as a dictionary.
:param retry_protocol_mismatches: Boolean to retry the transfer in case of protocol mismatch.
:returns: True if should retry it; False if no more retry.
"""
if is_intermediate_hop(req):
# This is an intermediate request in a multi-hop transfer. It must not be re-scheduled on its own.
# If needed, it will be re-scheduled via the creation of a new multi-hop transfer.
return False
if req['state'] == RequestState.SUBMITTING:
return True
if req['state'] == RequestState.NO_SOURCES or req['state'] == RequestState.ONLY_TAPE_SOURCES:
return False
# hardcoded for now - only requeue a couple of times
if req['retry_count'] is None or req['retry_count'] < 3:
if req['state'] == RequestState.MISMATCH_SCHEME:
return retry_protocol_mismatches
return True
return False
@METRICS.time_it
@transactional_session
def requeue_and_archive(request, source_ranking_update=True, retry_protocol_mismatches=False, *, session: "Session", logger=logging.log):
"""
Requeue and archive a failed request.
TODO: Multiple requeue.
:param request: Original request.
:param source_ranking_update Boolean. If True, the source ranking is decreased (making the sources less likely to be used)
:param session: Database session to use.
:param logger: Optional decorated logger that can be passed from the calling daemons or servers.
"""
# Probably not needed anymore
request_id = request['request_id']
new_req = get_request(request_id, session=session)
if new_req:
new_req['sources'] = get_sources(request_id, session=session)
archive_request(request_id, session=session)
if should_retry_request(new_req, retry_protocol_mismatches):
new_req['request_id'] = generate_uuid()
new_req['previous_attempt_id'] = request_id
if new_req['retry_count'] is None:
new_req['retry_count'] = 1
elif new_req['state'] != RequestState.SUBMITTING:
new_req['retry_count'] += 1
if source_ranking_update and new_req['sources']:
for i in range(len(new_req['sources'])):
if new_req['sources'][i]['is_using']:
if new_req['sources'][i]['ranking'] is None:
new_req['sources'][i]['ranking'] = -1
else:
new_req['sources'][i]['ranking'] -= 1
new_req['sources'][i]['is_using'] = False
new_req.pop('state', None)
queue_requests([new_req], session=session, logger=logger)
return new_req
else:
raise RequestNotFound
return None
@METRICS.count_it
@transactional_session
def queue_requests(requests, *, session: "Session", logger=logging.log):
"""
Submit transfer requests on destination RSEs for data identifiers.
:param requests: List of dictionaries containing request metadata.
:param session: Database session to use.
:param logger: Optional decorated logger that can be passed from the calling daemons or servers.
:returns: List of Request-IDs as 32 character hex strings.
"""
logger(logging.DEBUG, "queue requests")
request_clause = []
rses = {}
preparer_enabled = config_get_bool('conveyor', 'use_preparer', raise_exception=False, default=False)
for req in requests:
if isinstance(req['attributes'], str):
req['attributes'] = json.loads(req['attributes'] or '{}')
if isinstance(req['attributes'], str):
req['attributes'] = json.loads(req['attributes'] or '{}')
if req['request_type'] == RequestType.TRANSFER:
request_clause.append(and_(models.Request.scope == req['scope'],
models.Request.name == req['name'],
models.Request.dest_rse_id == req['dest_rse_id'],
models.Request.request_type == RequestType.TRANSFER))
if req['dest_rse_id'] not in rses:
rses[req['dest_rse_id']] = get_rse_name(req['dest_rse_id'], session=session)
# Check existing requests
existing_requests = []
if request_clause:
for requests_condition in chunks(request_clause, 1000):
stmt = select(
models.Request.scope,
models.Request.name,
models.Request.dest_rse_id
).with_hint(
models.Request, "INDEX(REQUESTS REQUESTS_SC_NA_RS_TY_UQ_IDX)", 'oracle'
).where(
or_(*requests_condition)
)
existing_requests.extend(session.execute(stmt))
new_requests, sources, messages = [], [], []
for request in requests:
dest_rse_name = get_rse_name(rse_id=request['dest_rse_id'], session=session)
if request['request_type'] == RequestType.TRANSFER and (request['scope'], request['name'], request['dest_rse_id']) in existing_requests:
logger(logging.WARNING, 'Request TYPE %s for DID %s:%s at RSE %s exists - ignoring' % (request['request_type'],
request['scope'],
request['name'],
dest_rse_name))
continue
def temp_serializer(obj):
if isinstance(obj, (InternalAccount, InternalScope)):
return obj.internal
raise TypeError('Could not serialise object %r' % obj)
if 'state' not in request:
request['state'] = RequestState.PREPARING if preparer_enabled else RequestState.QUEUED
new_request = {'request_type': request['request_type'],
'scope': request['scope'],
'name': request['name'],
'dest_rse_id': request['dest_rse_id'],
'source_rse_id': request.get('source_rse_id', None),
'attributes': json.dumps(request['attributes'], default=temp_serializer),
'state': request['state'],
'rule_id': request['rule_id'],
'activity': request['attributes']['activity'],
'bytes': request['attributes']['bytes'],
'md5': request['attributes']['md5'],
'adler32': request['attributes']['adler32'],
'account': request.get('account', None),
'priority': request['attributes'].get('priority', None),
'requested_at': request.get('requested_at', None),
'retry_count': request['retry_count']}
if 'transfertool' in request:
new_request['transfertool'] = request['transfertool']
if 'previous_attempt_id' in request and 'retry_count' in request:
new_request['previous_attempt_id'] = request['previous_attempt_id']
new_request['id'] = request['request_id']
else:
new_request['id'] = generate_uuid()
new_requests.append(new_request)
if 'sources' in request and request['sources']:
for source in request['sources']:
sources.append({'request_id': new_request['id'],
'scope': request['scope'],
'name': request['name'],
'rse_id': source['rse_id'],
'dest_rse_id': request['dest_rse_id'],
'ranking': source['ranking'],
'bytes': source['bytes'],
'url': source['url'],
'is_using': source['is_using']})
if request['request_type']:
transfer_status = '%s-%s' % (request['request_type'].name, request['state'].name)
else:
transfer_status = 'transfer-%s' % request['state'].name
transfer_status = transfer_status.lower()
payload = {'request-id': new_request['id'],
'request-type': request['request_type'].name.lower(),
'scope': request['scope'].external,
'name': request['name'],
'dst-rse-id': request['dest_rse_id'],
'dst-rse': dest_rse_name,
'state': request['state'].name.lower(),
'retry-count': request['retry_count'],
'rule-id': str(request['rule_id']),
'activity': request['attributes']['activity'],
'file-size': request['attributes']['bytes'],
'bytes': request['attributes']['bytes'],
'checksum-md5': request['attributes']['md5'],
'checksum-adler': request['attributes']['adler32'],
'queued_at': str(datetime.datetime.utcnow())}
messages.append({'event_type': transfer_status,
'payload': payload})
for requests_chunk in chunks(new_requests, 1000):
session.execute(insert(models.Request), requests_chunk)
for sources_chunk in chunks(sources, 1000):
session.execute(insert(models.Source), sources_chunk)
add_messages(messages, session=session)
return new_requests
@transactional_session
def list_and_mark_transfer_requests_and_source_replicas(
rse_collection: "RseCollection",
processed_by: Optional[str] = None,
processed_at_delay: int = 600,
total_workers: int = 0,
worker_number: int = 0,
partition_hash_var: Optional[str] = None,
limit: Optional[int] = None,
activity: Optional[str] = None,
older_than: Optional[datetime.datetime] = None,
rses: Optional[Sequence[str]] = None,
request_type: Optional[list[RequestType]] = None,
request_state: Optional[RequestState] = None,
required_source_rse_attrs: Optional[list[str]] = None,
ignore_availability: bool = False,
transfertool: Optional[str] = None,
*,
session: "Session",
) -> dict[str, RequestWithSources]:
"""
List requests with source replicas
:param rse_collection: the RSE collection being used
:param processed_by: the daemon/executable running this query
:param processed_at_delay: how many second to ignore a request if it's already being processed by the same daemon
:param total_workers: Number of total workers.
:param worker_number: Id of the executing worker.
:param partition_hash_var: The hash variable used for partitioning thread work
:param limit: Integer of requests to retrieve.
:param activity: Activity to be selected.
:param older_than: Only select requests older than this DateTime.
:param rses: List of rse_id to select requests.
:param request_type: Filter on the given request type.
:param request_state: Filter on the given request state
:param transfertool: The transfer tool as specified in rucio.cfg.
:param required_source_rse_attrs: Only select source RSEs having these attributes set
:param ignore_availability: Ignore blocklisted RSEs
:param session: Database session to use.
:returns: List of RequestWithSources objects.
"""
if partition_hash_var is None:
partition_hash_var = 'requests.id'
if request_state is None:
request_state = RequestState.QUEUED
if request_type is None:
request_type = [RequestType.TRANSFER]
sub_requests = select(
models.Request.id,
models.Request.request_type,
models.Request.rule_id,
models.Request.scope,
models.Request.name,
models.Request.md5,
models.Request.adler32,
models.Request.bytes,
models.Request.activity,
models.Request.attributes,
models.Request.previous_attempt_id,
models.Request.source_rse_id,
models.Request.dest_rse_id,
models.Request.retry_count,
models.Request.account,
models.Request.created_at,
models.Request.requested_at,
models.Request.priority,
models.Request.transfertool
).with_hint(
models.Request, "INDEX(REQUESTS REQUESTS_TYP_STA_UPD_IDX)", 'oracle'
).where(
models.Request.state == request_state,
models.Request.request_type.in_(request_type)
).join(
models.RSE,
models.RSE.id == models.Request.dest_rse_id
).where(
models.RSE.deleted == false()
).outerjoin(
models.TransferHop,
models.TransferHop.next_hop_request_id == models.Request.id
).where(
models.TransferHop.next_hop_request_id == null()
).order_by(
models.Request.created_at
)
if processed_by:
sub_requests = sub_requests.where(
or_(
models.Request.last_processed_by.is_(null()),
models.Request.last_processed_by != processed_by,
models.Request.last_processed_at < datetime.datetime.utcnow() - datetime.timedelta(seconds=processed_at_delay)
)
)
if not ignore_availability:
sub_requests = sub_requests.where(models.RSE.availability_write == true())
if isinstance(older_than, datetime.datetime):
sub_requests = sub_requests.where(models.Request.requested_at < older_than)
if activity:
sub_requests = sub_requests.where(models.Request.activity == activity)
# if a transfertool is specified make sure to filter for those requests and apply related index
if transfertool:
sub_requests = sub_requests.where(models.Request.transfertool == transfertool)
sub_requests = sub_requests.with_hint(models.Request, "INDEX(REQUESTS REQUESTS_TYP_STA_TRA_ACT_IDX)", 'oracle')
else:
sub_requests = sub_requests.with_hint(models.Request, "INDEX(REQUESTS REQUESTS_TYP_STA_UPD_IDX)", 'oracle')
if rses:
temp_table_cls = temp_table_mngr(session).create_id_table()
session.execute(insert(temp_table_cls), [{'id': rse_id} for rse_id in rses])
sub_requests = sub_requests.join(temp_table_cls, temp_table_cls.id == models.RSE.id)
sub_requests = filter_thread_work(session=session, query=sub_requests, total_threads=total_workers, thread_id=worker_number, hash_variable=partition_hash_var)
if limit:
sub_requests = sub_requests.limit(limit)
sub_requests = sub_requests.subquery()
stmt = select(
sub_requests.c.id,
sub_requests.c.request_type,
sub_requests.c.rule_id,
sub_requests.c.scope,
sub_requests.c.name,
sub_requests.c.md5,
sub_requests.c.adler32,
sub_requests.c.bytes,
sub_requests.c.activity,
sub_requests.c.attributes,
sub_requests.c.previous_attempt_id,
sub_requests.c.source_rse_id,
sub_requests.c.dest_rse_id,
sub_requests.c.account,
sub_requests.c.retry_count,
sub_requests.c.priority,
sub_requests.c.transfertool,
sub_requests.c.requested_at,
models.RSE.id.label("replica_rse_id"),
models.RSE.rse.label("replica_rse_name"),
models.RSEFileAssociation.path,
models.Source.ranking.label("source_ranking"),
models.Source.url.label("source_url"),
models.Distance.distance
).order_by(
sub_requests.c.created_at
).outerjoin(
models.RSEFileAssociation,
and_(sub_requests.c.scope == models.RSEFileAssociation.scope,
sub_requests.c.name == models.RSEFileAssociation.name,
models.RSEFileAssociation.state == ReplicaState.AVAILABLE,
sub_requests.c.dest_rse_id != models.RSEFileAssociation.rse_id)
).with_hint(
models.RSEFileAssociation, "INDEX(REPLICAS REPLICAS_PK)", 'oracle'
).outerjoin(
models.RSE,
and_(models.RSE.id == models.RSEFileAssociation.rse_id,
models.RSE.deleted == false())
).outerjoin(
models.Source,
and_(sub_requests.c.id == models.Source.request_id,
models.RSE.id == models.Source.rse_id)
).with_hint(
models.Source, "INDEX(SOURCES SOURCES_PK)", 'oracle'
).outerjoin(
models.Distance,
and_(sub_requests.c.dest_rse_id == models.Distance.dest_rse_id,
models.RSEFileAssociation.rse_id == models.Distance.src_rse_id)
).with_hint(
models.Distance, "INDEX(DISTANCES DISTANCES_PK)", 'oracle'
)
for attribute in required_source_rse_attrs or ():
rse_attr_alias = aliased(models.RSEAttrAssociation)
stmt = stmt.where(
exists(
select(
1
).where(
rse_attr_alias.rse_id == models.RSE.id,
rse_attr_alias.key == attribute
)
)
)
requests_by_id = {}
for (request_id, req_type, rule_id, scope, name, md5, adler32, byte_count, activity, attributes, previous_attempt_id, source_rse_id, dest_rse_id, account, retry_count,
priority, transfertool, requested_at, replica_rse_id, replica_rse_name, file_path, source_ranking, source_url, distance) in session.execute(stmt):
request = requests_by_id.get(request_id)
if not request:
request = RequestWithSources(id_=request_id, request_type=req_type, rule_id=rule_id, scope=scope, name=name,
md5=md5, adler32=adler32, byte_count=byte_count, activity=activity, attributes=attributes,
previous_attempt_id=previous_attempt_id, dest_rse=rse_collection[dest_rse_id],
account=account, retry_count=retry_count, priority=priority, transfertool=transfertool,
requested_at=requested_at)
requests_by_id[request_id] = request
# if STAGEIN and destination RSE is QoS make sure the source is included
if request.request_type == RequestType.STAGEIN and get_rse_attribute(rse_id=dest_rse_id, key='staging_required', session=session):
source = RequestSource(rse=rse_collection[dest_rse_id])
request.sources.append(source)
if replica_rse_id is not None:
replica_rse = rse_collection[replica_rse_id]
replica_rse.name = replica_rse_name
source = RequestSource(rse=replica_rse, file_path=file_path,
ranking=source_ranking, distance=distance, url=source_url)
request.sources.append(source)
if source_rse_id == replica_rse_id:
request.requested_source = source
if processed_by:
for chunk in chunks(requests_by_id, 100):
stmt = update(
models.Request
).where(
models.Request.id.in_(chunk)
).execution_options(
synchronize_session=False
).values(
{
models.Request.last_processed_by: processed_by,
models.Request.last_processed_at: datetime.datetime.now(),
}
)
session.execute(stmt)
return requests_by_id
@read_session
def fetch_paths(request_id, *, session: "Session"):
"""
Find the paths for which the provided request is a constituent hop.
Returns a dict: {initial_request_id1: path1, ...}. Each path is an ordered list of request_ids.
"""
transfer_hop_alias = aliased(models.TransferHop)
stmt = select(
models.TransferHop,
).join(
transfer_hop_alias,
and_(
transfer_hop_alias.initial_request_id == models.TransferHop.initial_request_id,
or_(transfer_hop_alias.request_id == request_id,
transfer_hop_alias.initial_request_id == request_id),
)
)
parents_by_initial_request = {}
for hop, in session.execute(stmt):
parents_by_initial_request.setdefault(hop.initial_request_id, {})[hop.next_hop_request_id] = hop.request_id
paths = {}
for initial_request_id, parents in parents_by_initial_request.items():
path = []
cur_request = initial_request_id
path.append(cur_request)
while parents.get(cur_request):
cur_request = parents[cur_request]
path.append(cur_request)
paths[initial_request_id] = list(reversed(path))
return paths
@METRICS.time_it
@transactional_session
def get_and_mark_next(
rse_collection: "RseCollection",
request_type,
state,
processed_by: Optional[str] = None,
processed_at_delay: int = 600,
limit: int = 100,
older_than: "Optional[datetime.datetime]" = None,
rse_id: "Optional[str]" = None,
activity: "Optional[str]" = None,
total_workers: int = 0,
worker_number: int = 0,
mode_all=False,
hash_variable='id',
activity_shares=None,
include_dependent=True,
transfertool=None,
*,
session: "Session"
):
"""
Retrieve the next requests matching the request type and state.
Workers are balanced via hashing to reduce concurrency on database.
:param rse_collection: the RSE collection being used
:param request_type: Type of the request as a string or list of strings.
:param state: State of the request as a string or list of strings.
:param processed_by: the daemon/executable running this query
:param processed_at_delay: how many second to ignore a request if it's already being processed by the same daemon
:param limit: Integer of requests to retrieve.
:param older_than: Only select requests older than this DateTime.
:param rse_id: The RSE to filter on.
:param activity: The activity to filter on.
:param total_workers: Number of total workers.
:param worker_number: Id of the executing worker.
:param mode_all: If set to True the function returns everything, if set to False returns list of dictionaries {'request_id': x, 'external_host': y, 'external_id': z}.
:param hash_variable: The variable to use to perform the partitioning. By default it uses the request id.
:param activity_shares: Activity shares dictionary, with number of requests
:param include_dependent: If true, includes transfers which have a previous hop dependency on other transfers
:param transfertool: The transfer tool as specified in rucio.cfg.
:param session: Database session to use.
:returns: Request as a dictionary.
"""
request_type_metric_label = '.'.join(a.name for a in request_type) if isinstance(request_type, list) else request_type.name
state_metric_label = '.'.join(s.name for s in state) if isinstance(state, list) else state.name
METRICS.counter('get_next.requests.{request_type}.{state}').labels(request_type=request_type_metric_label, state=state_metric_label).inc()
# lists of one element are not allowed by SQLA, so just duplicate the item
if type(request_type) is not list:
request_type = [request_type, request_type]
elif len(request_type) == 1:
request_type = [request_type[0], request_type[0]]
if type(state) is not list:
state = [state, state]
elif len(state) == 1:
state = [state[0], state[0]]
result = []
if not activity_shares:
activity_shares = [None]
for share in activity_shares:
query = select(
models.Request.id
).where(
models.Request.state.in_(state),
models.Request.request_type.in_(request_type)
).order_by(
asc(models.Request.updated_at)
)
if processed_by:
query = query.where(
or_(
models.Request.last_processed_by.is_(null()),
models.Request.last_processed_by != processed_by,
models.Request.last_processed_at < datetime.datetime.utcnow() - datetime.timedelta(seconds=processed_at_delay)
)
)
if transfertool:
query = query.with_hint(
models.Request, "INDEX(REQUESTS REQUESTS_TYP_STA_TRA_ACT_IDX)", 'oracle'
).where(
models.Request.transfertool == transfertool
)
else:
query = query.with_hint(
models.Request, "INDEX(REQUESTS REQUESTS_TYP_STA_UPD_IDX)", 'oracle'
)
if not include_dependent:
# filter out transfers which depend on some other "previous hop" requests.
# In particular, this is used to avoid multiple finishers trying to archive different
# transfers from the same path and thus having concurrent deletion of same rows from
# the transfer_hop table.
query = query.outerjoin(
models.TransferHop,
models.TransferHop.next_hop_request_id == models.Request.id
).where(
models.TransferHop.next_hop_request_id == null()
)
if isinstance(older_than, datetime.datetime):
query = query.filter(models.Request.updated_at < older_than)
if rse_id:
query = query.filter(models.Request.dest_rse_id == rse_id)
if share:
query = query.filter(models.Request.activity == share)
elif activity:
query = query.filter(models.Request.activity == activity)
query = filter_thread_work(session=session, query=query, total_threads=total_workers, thread_id=worker_number, hash_variable=hash_variable)
if share:
query = query.limit(activity_shares[share])
else:
query = query.limit(limit)
if session.bind.dialect.name == 'oracle':
query = select(
models.Request
).where(
models.Request.id.in_(query)
).with_for_update(
skip_locked=True
)
else:
query = query.with_only_columns(
models.Request
).with_for_update(
skip_locked=True,
of=models.Request.last_processed_by
)
query_result = session.execute(query).scalars()
if query_result:
if mode_all:
for res in query_result:
res_dict = res.to_dict()
res_dict['request_id'] = res_dict['id']
res_dict['attributes'] = json.loads(str(res_dict['attributes'] or '{}'))
dst_id = res_dict['dest_rse_id']
src_id = res_dict['source_rse_id']
res_dict['dst_rse'] = rse_collection[dst_id].ensure_loaded(load_name=True, load_attributes=True)
res_dict['src_rse'] = rse_collection[src_id].ensure_loaded(load_name=True, load_attributes=True) if src_id is not None else None
result.append(res_dict)
else:
for res in query_result:
result.append({'request_id': res.id, 'external_host': res.external_host, 'external_id': res.external_id})
request_ids = {r['request_id'] for r in result}
if processed_by and request_ids:
for chunk in chunks(request_ids, 100):
stmt = update(
models.Request
).where(
models.Request.id.in_(chunk)
).execution_options(
synchronize_session=False
).values(
{
models.Request.last_processed_by: processed_by,
models.Request.last_processed_at: datetime.datetime.now(),
}
)
session.execute(stmt)
return result
@transactional_session
def update_request(
request_id: str,
state: Optional[RequestState] = None,
transferred_at: Optional[datetime.datetime] = None,
started_at: Optional[datetime.datetime] = None,
staging_started_at: Optional[datetime.datetime] = None,
staging_finished_at: Optional[datetime.datetime] = None,
source_rse_id: Optional[str] = None,
err_msg: Optional[str] = None,
attributes: Optional[dict[str, str]] = None,
priority: Optional[int] = None,
transfertool: Optional[str] = None,
*,
raise_on_missing: bool = False,
session: "Session",
):
rowcount = 0
try:
update_items: dict[Any, Any] = {
models.Request.updated_at: datetime.datetime.utcnow()
}
if state is not None:
update_items[models.Request.state] = state
if transferred_at is not None:
update_items[models.Request.transferred_at] = transferred_at
if started_at is not None:
update_items[models.Request.started_at] = started_at
if staging_started_at is not None:
update_items[models.Request.staging_started_at] = staging_started_at
if staging_finished_at is not None:
update_items[models.Request.staging_finished_at] = staging_finished_at
if source_rse_id is not None:
update_items[models.Request.source_rse_id] = source_rse_id
if err_msg is not None:
update_items[models.Request.err_msg] = err_msg
if attributes is not None:
update_items[models.Request.attributes] = json.dumps(attributes)
if priority is not None:
update_items[models.Request.priority] = priority
if transfertool is not None:
update_items[models.Request.transfertool] = transfertool
stmt = update(
models.Request
).where(
models.Request.id == request_id
).execution_options(
synchronize_session=False
).values(
update_items
)
rowcount = session.execute(stmt).rowcount
except IntegrityError as error:
raise RucioException(error.args)
if not rowcount and raise_on_missing:
raise UnsupportedOperation("Request %s state cannot be updated." % request_id)
if rowcount:
return True
return False
@METRICS.count_it
@transactional_session
def transition_request_state(
request_id: str,
state: Optional[RequestState] = None,
external_id: Optional[str] = None,
transferred_at: Optional[datetime.datetime] = None,
started_at: Optional[datetime.datetime] = None,
staging_started_at: Optional[datetime.datetime] = None,
staging_finished_at: Optional[datetime.datetime] = None,
source_rse_id: Optional[str] = None,
err_msg: Optional[str] = None,
attributes: Optional[dict[str, str]] = None,
*,
request: "Optional[dict[str, Any]]" = None,
session: "Session",
logger=logging.log
) -> bool:
"""
Update the request if its state changed. Return a boolean showing if the request was actually updated or not.
"""
# TODO: Should this be a private method?
if request is None:
request = get_request(request_id, session=session)
if not request:
# The request was deleted in the meantime. Ignore it.
logger(logging.WARNING, "Request %s not found. Cannot set its state to %s", request_id, state)
return False
if request['state'] == state:
logger(logging.INFO, "Request %s state is already %s. Will skip the update.", request_id, state)
return False
if state in [RequestState.FAILED, RequestState.DONE, RequestState.LOST] and (request["external_id"] != external_id):
logger(logging.ERROR, "Request %s should not be updated to 'Failed' or 'Done' without external transfer_id" % request_id)
return False
update_request(
request_id=request_id,
state=state,
transferred_at=transferred_at,
started_at=started_at,
staging_started_at=staging_started_at,
staging_finished_at=staging_finished_at,
source_rse_id=source_rse_id,
err_msg=err_msg,
attributes=attributes,
raise_on_missing=True,
session=session,
)
return True
@METRICS.count_it
@transactional_session
def transition_requests_state_if_possible(request_ids, new_state, *, session: "Session", logger=logging.log):
"""
Bulk update the state of requests. Skips silently if the request_id does not exist.
:param request_ids: List of (Request-ID as a 32 character hex string).
:param new_state: New state as string.
:param session: Database session to use.
:param logger: Optional decorated logger that can be passed from the calling daemons or servers.
"""
try:
for request_id in request_ids:
try:
transition_request_state(request_id, new_state, session=session, logger=logger)
except UnsupportedOperation:
continue
except IntegrityError as error: