Igor Mandrichenko
Igor Mandrichenko
I looked at the new temp table-based core.list_replicas() code. It seems to be quite complicated. I am going to de-assign myself from this issue, hoping one of the two things...
PR: https://github.com/rucio/rucio/pull/5428
Waiting for PR https://github.com/rucio/rucio/pull/5798 to be merged in
Apparently using api.declare_bad_file_replicas() method is not correct way to declare missing replicas. The underlying core.declare_bad_file_replicas() checks whether the replica is already known to Necromancer as bad by looking it up...
I am going to assume that it is OK to force updating missing replica state in `replicas` table to "bad" ignoring its state in the `bad_replicas` table. I am going...
The PR to add the "force" flag is in: https://github.com/rucio/rucio/pull/5864 Please review. Thanks !
I was trying to propose a slightly different solution. Instead of trying to convert one text representation format to another, parse the text representation of DNs from the client and...
Does Oracle support JSON as native data type ? If it was Postgres, I would store JSON representation of the parsed DN in the database as jsonb data.
Currently, there are 2 ways to query replicas in datasets: 1. Query multiple datasets and return basic info about each replica found in these datasets (file scope, path, rse_id, etc...)...
After Martin's comment, do we still want this change done ?