usage: rucio-replica-recoverer [-h] [--nattempts NATTEMPTS]
[--younger-than YOUNGER_THAN]
[--vos VOS [VOS ...]] [--run-once]
[--limit-suspicious-files-on-rse LIMIT_SUSPICIOUS_FILES_ON_RSE]
[--json-file-name JSON_FILE_NAME]
[--sleep-time SLEEP_TIME] [--active-mode]
Replica-Recoverer is a daemon that declares suspicious replicas
that are available on other RSE as bad. Consequently, automatic
replica recovery is triggered via necromancer daemon,
which creates a rule for such bad replicas.
optional arguments:
-h, --help show this help message and exit
--nattempts NATTEMPTS
Minimum count of suspicious file replica appearance in
bad_replicas table. Default value is 5.
--younger-than YOUNGER_THAN
Consider all file replicas logged in bad_replicas
table since speicified number of younger-than days.
Default value is 5.
--vos VOS [VOS ...] Optional list of VOs to consider. Only used in multi-
VO mode.
--run-once One iteration only.
--limit-suspicious-files-on-rse LIMIT_SUSPICIOUS_FILES_ON_RSE
Maximum number of suspicious replicas on an RSE before
that RSE is considered problematic and the suspicious
replicas on that RSE are declared
"TEMPORARY_UNAVAILABLE". Default value is 5.
--json-file-name JSON_FILE_NAME
Name of the json file that that contains the policies
which will be used by the suspicious replica
recoverer.
--sleep-time SLEEP_TIME
Concurrency control: Thread sleep time after each
chunk of work.
--active-mode If NOT specified, the daemon will run without taking
any actions on any files. In either case, the log file
will be produced normally.
Preparing RSEs and DIDs for testing this daemon from rucio docker:
-------------------------------------------------------------------
$ sudo docker exec -it dev_rucio_1 /bin/bash
Adding the RSEs to rucio DB:
$ rucio-admin rse add MOCK_SUSPICIOUS
$ rucio-admin rse set-attribute --rse MOCK_SUSPICIOUS --key backend_type --value file
$ rucio-admin rse set-attribute --rse MOCK_SUSPICIOUS --key storage_usage_tool --value 'rucio.rse.protocols.posix.Default.getSpace'
$ rucio-admin rse add-protocol --hostname localhost --scheme file --prefix '/tmp/rucio_rse1/' --space-token 'ATLASDATADISK1' --web-service-path '/srm/managerv2?SFN=' --domain-json '{ "lan": { "read": 1, "write": 1, "delete": 1 }, "wan": { "read": 1, "write": 1, "delete": 1}}' --impl 'rucio.rse.protocols.posix.Default' MOCK_SUSPICIOUS
$ rucio-admin rse info MOCK_SUSPICIOUS
$ rucio-admin rse add MOCK_RECOVERY
$ rucio-admin rse set-attribute --rse MOCK_RECOVERY --key backend_type --value POSIX
$ rucio-admin rse set-attribute --rse MOCK_RECOVERY --key storage_usage_tool --value 'rucio.rse.protocols.posix.Default.getSpace'
$ rucio-admin rse add-protocol --hostname localhost --scheme file --prefix '/tmp/rucio_rse2/' --space-token 'ATLASDATADISK2' --web-service-path '/srm/managerv2?SFN=' --domain-json '{ "lan": { "read": 1, "write": 1, "delete": 1 }, "wan": { "read": 1, "write": 1, "delete": 1}}' --impl 'rucio.rse.protocols.posix.Default' MOCK_RECOVERY
$ rucio-admin rse info MOCK_RECOVERY
For testing, we create the following files:
-------------------------------------------------------------------
Name MOCK_RECOVERY MOCK_SUSPICIOUS BAD (on MOCK_SUSPICIOUS) SUSPICIOUS (on MOCK_SUSPICIOUS)
file_available_suspicious yes yes no yes
file_available_suspicious_and_bad yes yes yes yes
file_notavailable_suspicious unavailable yes no yes
Only file_available_suspicious should be the one on which the daemon takes action and declares it as bad.
$ id0=`uuidgen`
$ id1=`uuidgen`
$ id2=`uuidgen`
$ id3=`uuidgen`
$ echo "file available on MOCK_RECOVERY and decalred suspicious on MOCK_SUSPICIOUS (11 times)" > /tmp/file_available_suspicious'_'$id1
$ echo "file available on MOCK_RECOVERY and declared suspicious on MOCK_SUSPICIOUS (11 times) and 1 time bad/deleted/lost on MOCK_SUSPICIOUS" > /tmp/file_available_suspicious_and_bad'_'$id2
$ echo "file declared as unavailable on MOCK_RECOVERY and declared as suspicious 11 times on MOCK_SUSPICIOUS" > /tmp/file_notavailable_suspicious'_'$id3
Uploading the files created above to rucio:
-------------------------------------------------------------------
rucio add-dataset mock:dataset_of_suspicious_replicas'_'$id0
Added mock:dataset_of_suspicious_replicas_2ba45524-860b-43f9-a601-6ccec2c46778
$ rucio add-rule mock:dataset_of_suspicious_replicas'_'$id0 1 MOCK_SUSPICIOUS
$ rucio add-rule --source-replica-expression MOCK_SUSPICIOUS mock:dataset_of_suspicious_replicas'_'$id0 1 MOCK_RECOVERY
$ rucio list-rules mock:dataset_of_suspicious_replicas'_'$id0
ID ACCOUNT SCOPE:NAME STATE[OK/REPL/STUCK] RSE_EXPRESSION COPIES EXPIRES (UTC) CREATED (UTC)
-------------------------------- --------- ------------------------------------------------------------------------ ---------------------- ---------------- -------- --------------- -------------------
2a1a078b66ca4e209cc20a5826125334 root mock:dataset_of_suspicious_replicas_2ba45524-860b-43f9-a601-6ccec2c46778 OK[0/0/0] MOCK_RECOVERY 1 2019-02-19 14:12:30
8c15d2f8e94a459a86a488055a10d068 root mock:dataset_of_suspicious_replicas_2ba45524-860b-43f9-a601-6ccec2c46778 OK[0/0/0] MOCK_SUSPICIOUS 1 2019-02-19 14:12:28
$ rucio upload --scope mock --rse MOCK_SUSPICIOUS --name file_available_suspicious'_'$id1 /tmp/file_available_suspicious'_'$id1
$ rucio upload --scope mock --rse MOCK_SUSPICIOUS --name file_available_suspicious_and_bad'_'$id2 /tmp/file_available_suspicious_and_bad'_'$id2
$ rucio upload --scope mock --rse MOCK_SUSPICIOUS --name file_notavailable_suspicious'_'$id3 /tmp/file_notavailable_suspicious'_'$id3
$ rucio upload --scope mock --rse MOCK_RECOVERY --name file_available_suspicious'_'$id1 /tmp/file_available_suspicious'_'$id1
$ rucio upload --scope mock --rse MOCK_RECOVERY --name file_available_suspicious_and_bad'_'$id2 /tmp/file_available_suspicious_and_bad'_'$id2
$ rucio upload --scope mock --rse MOCK_RECOVERY --name file_notavailable_suspicious'_'$id3 /tmp/file_notavailable_suspicious'_'$id3
[...]
$ rucio attach mock:dataset_of_suspicious_replicas'_'$id0 mock:file_available_suspicious'_'$id1
$ rucio attach mock:dataset_of_suspicious_replicas'_'$id0 mock:file_available_suspicious_and_bad'_'$id2
$ rucio attach mock:dataset_of_suspicious_replicas'_'$id0 mock:file_notavailable_suspicious'_'$id3
[...]
$ rucio list-file-replicas mock:file_available_suspicious'_'$id1
+---------+----------------------------------------------------------------+------------+-----------+------------------------------------------------------------------------------------------------------------------------------+
| SCOPE | NAME | FILESIZE | ADLER32 | RSE: REPLICA |
|---------+----------------------------------------------------------------+------------+-----------+------------------------------------------------------------------------------------------------------------------------------|
| mock | file_available_suspicious_5180be3e-4ebc-4c34-b528-efbfd09f067e | 87.000 B | 206b1c91 | MOCK_SUSPICIOUS: file://localhost:0/tmp/rucio_rse1/mock/1f/6b/file_available_suspicious_5180be3e-4ebc-4c34-b528-efbfd09f067e |
| mock | file_available_suspicious_5180be3e-4ebc-4c34-b528-efbfd09f067e | 87.000 B | 206b1c91 | MOCK_RECOVERY: file://localhost:0/tmp/rucio_rse2/mock/1f/6b/file_available_suspicious_5180be3e-4ebc-4c34-b528-efbfd09f067e |
+---------+----------------------------------------------------------------+------------+-----------+------------------------------------------------------------------------------------------------------------------------------+
$ rucio list-file-replicas mock:file_available_suspicious_and_bad'_'$id2
+---------+------------------------------------------------------------------------+------------+-----------+--------------------------------------------------------------------------------------------------------------------------------------+
| SCOPE | NAME | FILESIZE | ADLER32 | RSE: REPLICA |
|---------+------------------------------------------------------------------------+------------+-----------+--------------------------------------------------------------------------------------------------------------------------------------|
| mock | file_available_suspicious_and_bad_46964411-95d4-46c4-a973-72c045195835 | 134.000 B | dfdf2bff | MOCK_SUSPICIOUS: file://localhost:0/tmp/rucio_rse1/mock/1d/2d/file_available_suspicious_and_bad_46964411-95d4-46c4-a973-72c045195835 |
| mock | file_available_suspicious_and_bad_46964411-95d4-46c4-a973-72c045195835 | 134.000 B | dfdf2bff | MOCK_RECOVERY: file://localhost:0/tmp/rucio_rse2/mock/1d/2d/file_available_suspicious_and_bad_46964411-95d4-46c4-a973-72c045195835 |
+---------+------------------------------------------------------------------------+------------+-----------+--------------------------------------------------------------------------------------------------------------------------------------+
$ rucio list-file-replicas mock:file_notavailable_suspicious'_'$id3
+---------+-------------------------------------------------------------------+------------+-----------+---------------------------------------------------------------------------------------------------------------------------------+
| SCOPE | NAME | FILESIZE | ADLER32 | RSE: REPLICA |
|---------+-------------------------------------------------------------------+------------+-----------+---------------------------------------------------------------------------------------------------------------------------------|
| mock | file_notavailable_suspicious_6157f589-80db-492c-acdd-ef5f0c45112f | 101.000 B | 0c14223f | MOCK_SUSPICIOUS: file://localhost:0/tmp/rucio_rse1/mock/7d/a6/file_notavailable_suspicious_6157f589-80db-492c-acdd-ef5f0c45112f |
| mock | file_notavailable_suspicious_6157f589-80db-492c-acdd-ef5f0c45112f | 101.000 B | 0c14223f | MOCK_RECOVERY: file://localhost:0/tmp/rucio_rse2/mock/7d/a6/file_notavailable_suspicious_6157f589-80db-492c-acdd-ef5f0c45112f |
+---------+-------------------------------------------------------------------+------------+-----------+---------------------------------------------------------------------------------------------------------------------------------+
Modifying the file statuses in the DB:
--------------------------------------
$ python
# the paths below point to MOCK_SUSPICIOUS RSE (.../rucio_rse1)
$$ file1 = ['file://localhost:0/tmp/rucio_rse1/mock/1f/6b/file_available_suspicious_5180be3e-4ebc-4c34-b528-efbfd09f067e',]
$$ file2 = ['file://localhost:0/tmp/rucio_rse1/mock/1d/2d/file_available_suspicious_and_bad_46964411-95d4-46c4-a973-72c045195835',]
$$ file3 = ['file://localhost:0/tmp/rucio_rse1/mock/7d/a6/file_notavailable_suspicious_6157f589-80db-492c-acdd-ef5f0c45112f' ]
$$ from rucio.client.replicaclient import ReplicaClient
$$ replica_client = ReplicaClient()
$$ import time
$$ for i in range(11):
replica_client.declare_suspicious_file_replicas(file1, 'This is a good reason')
replica_client.declare_suspicious_file_replicas(file2, 'This is a good reason')
replica_client.declare_suspicious_file_replicas(file3, 'This is a good reason')
time.sleep(1)
# Declaring file2 bad on MOCK_SUSPICIOUS:
---------------------------------------
$$ replica_client.declare_bad_file_replicas(file2, 'This is a good reason')
# Update replica state of 'file_notavailable_suspicious'_'$id1' on MOCK_RECOVERY to 'UNAVAILABLE'
# (change the file name below according to the info from rucio !):
-----------------------------------------------------------------------------------------------
$$ replica_client.update_replicas_states('MOCK_RECOVERY', [{'scope':'mock', 'name':'file_notavailable_suspicious_6157f589-80db-492c-acdd-ef5f0c45112f', 'state':'U'}])
# Checking the results of the file status changes:
------------------------------------------------
$$ from rucio.core.replica import get_suspicious_files
$$ from datetime import datetime, timedelta
$$ from_date = datetime.now() - timedelta(days=3)
$$ from rucio.core.replica import list_bad_replicas_status
$$ get_suspicious_files('MOCK_SUSPICIOUS',from_date,10)
[{'created_at': datetime.datetime(2019, 2, 19, 14, 12, 56), 'scope': 'mock', 'cnt': 11L, 'name': 'file_notavailable_suspicious_6157f589-80db-492c-acdd-ef5f0c45112f', 'rse': 'MOCK_SUSPICIOUS'},
{'created_at': datetime.datetime(2019, 2, 19, 14, 12, 48), 'scope': 'mock', 'cnt': 11L, 'name': 'file_available_suspicious_5180be3e-4ebc-4c34-b528-efbfd09f067e', 'rse': 'MOCK_SUSPICIOUS'}]
$$ list_bad_replicas_status(rse='MOCK_SUSPICIOUS', younger_than=from_date)
[{'name': 'file_available_suspicious_and_bad_46964411-95d4-46c4-a973-72c045195835', 'rse': 'MOCK_SUSPICIOUS', 'created_at': datetime.datetime(2019, 2, 19, 14, 18, 33), 'updated_at': datetime.datetime(2019, 2, 19, 14, 18, 33), 'state': BAD, 'scope': 'mock'}]
$$ exit()
Run the daemon:
---------------
$ python bin/rucio-replica-recoverer --run-once --rse-expression='MOCK_SUSPICIOUS'
Terminal output:
----------------
2019-02-19 14:39:24,114 709 INFO replica_recoverer[0/0]: ready to query replicas at RSEs like *MOCK*, declared as suspicious in the last 3 days at least 10 times and which are available on other RSEs.
2019-02-19 14:39:24,124 709 INFO replica_recoverer[0/0]: suspicious replica query took 0.0101511478424 seconds, total of 1 replicas were found. [{'scope': 'mock', 'cnt': 11L, 'name': 'file_available_suspicious_5180be3e-4ebc-4c34-b528-efbfd09f067e', 'rse': 'MOCK_SUSPICIOUS'}]
2019-02-19 14:39:24,125 709 INFO replica_recoverer[0/0]: looking for replica surls.
2019-02-19 14:39:24,160 709 INFO replica_recoverer[0/0]: found 1/1 surls (took 0.035572052002 seconds) - declaring them as bad replicas now.
2019-02-19 14:39:24,160 709 INFO replica_recoverer[0/0]: ready to declare 1 bad replica(s) on MOCK_SUSPICIOUS: ['file://localhost:0/tmp/rucio_rse1/mock/1f/6b/file_available_suspicious_5180be3e-4ebc-4c34-b528-efbfd09f067e'].
2019-02-19 14:39:24,188 709 INFO replica_recoverer[0/0]: finished declaring bad replicas on MOCK_SUSPICIOUS.
2019-02-19 14:39:24,192 709 INFO replica_recoverer[0/0]: graceful stop done
# Checking the results of the file status changes:
------------------------------------------------
$ python
$$ from rucio.core.replica import get_suspicious_files
$$ from datetime import datetime, timedelta
$$ from_date = datetime.now() - timedelta(days=3)
$$ from rucio.core.replica import list_bad_replicas_status
$$ from rucio.common.constants import SuspiciousAvailability
$$ get_suspicious_files('MOCK_SUSPICIOUS',younger_than=from_date, nattempts=10, is_suspicious=True, available_elsewhere=SuspiciousAvailability["EXIST_COPIES"].value)
>>> get_suspicious_files('MOCK_SUSPICIOUS',younger_than=from_date, nattempts=10, is_suspicious=True, available_elsewhere=SuspiciousAvailability["EXIST_COPIES"].value)
[]
$$ list_bad_replicas_status(rse='MOCK_SUSPICIOUS', younger_than=from_date)
[{'name': 'file_available_suspicious_5180be3e-4ebc-4c34-b528-efbfd09f067e', 'rse': 'MOCK_SUSPICIOUS', 'created_at': datetime.datetime(2019, 2, 19, 14, 39, 24), 'updated_at': datetime.datetime(2019, 2, 19, 14, 39, 24), 'state': BAD, 'scope': 'mock'},
{'name': 'file_available_suspicious_and_bad_46964411-95d4-46c4-a973-72c045195835', 'rse': 'MOCK_SUSPICIOUS', 'created_at': datetime.datetime(2019, 2, 19, 14, 18, 33), 'updated_at': datetime.datetime(2019, 2, 19, 14, 18, 33), 'state': BAD, 'scope': 'mock'}]
$$ exit()
When run in multi-VO mode, by default the daemon will run on RSEs from all VOs::
$ rucio-replica-recoverer --run-once
2020-07-28 15:15:14,151 5461 INFO replica_recoverer: This instance will work on VOs: def, abc, xyz, 123
By using the ``--vos`` argument only the VO or VOs specified will be affected::
$ rucio-replica-recoverer --run-once --vos abc xyz
2020-07-28 15:16:36,066 5474 INFO replica_recoverer: This instance will work on VOs: abc, xyz
Note that attempting the use the ``--vos`` argument when in single-VO mode will have no affect::
$ rucio-replica-recoverer --run-once --vos abc xyz
2020-07-28 15:21:33,349 5488 WARNING Ignoring argument vos, this is only applicable in a multi-VO setup.