cserf
cserf
After a bit of investigation, I think that the problem is in the `detach_did` workflow. The following lines : https://github.com/rucio/rucio/blob/1f10818974abd07681b7f3b74c430ca20ccb41b4/lib/rucio/core/rule.py#L2458-L2468 do remove the dataset locks from the children if the...
Reassigning the ticket to @bari12 :D
Please don't do it for Hermes2. I have a jumbo feature PR open, and the merge might be complicated.
It might overlap with : https://github.com/rucio/rucio/issues/5839 Have you already started working on it ? If yes, I will wait before working on the other ticket
I will just work on the following ones: ``` - rucio_tests/test_rse_protocol_ssh.py - rucio_tests/test_redirect.py - rucio_tests/test_credential.py - rucio_tests/test_bad_replica.py - rucio_tests/test_bb8.py - rucio_tests/test_rse_lfn2path.py - rucio_tests/test_s3.py - rucio_tests/test_rule.py - rucio_tests/test_multi_vo.py - rucio_tests/test_naming_convention.py -...
The problem is not the `necromancer` but `minos`. The datasets containing the files are not resolved. As you mentioned, resolving them would be an expensive operation which would slow down...
Doing extra check on the necromancer. There are 2 methods used by the necromancer depending if the file is recoverable or not : - `update_rules_for_bad_replica` in case the replica is...
Setting a rule is needed to protect the replicas. If it is not done, the replica is created with a default tombstone and can disappear immediately in case the reaper...
Rebased. About defining the table in `etc/sql/oracle/schema.sql`, is it needed ? I think this file is only used by ATLAS. Moreover `did_meta` table is not defined in this file.