rcarpa

Any updates on my question ?

I'm happy that did_type cannot be set, I didn't know about that. This means that this is a hack we can actually get rid of. I consider that we must...

I just performed this test ```python from rucio.core.config import get as cconfig_get, set as cconfig_set from rucio.client import Client def test_many_replicas(replica_client, mock_scope, core_config_mock): files = [{'scope': mock_scope.external, 'name': 'file_%s' %...

resolve_dids is not the problem here: ``` resolve_dids took 0:00:00.528797 Took with use_temp_table=False: 0:00:19.360680 ``` Has to be investigated, but it may be that the postgresql optimizer does a bad...

This is correct and should be fixed. Nevertheless, rucio relies very heavily on caching in other parts of the code and you'll experience bad performance in many other places too...

I'll pick it up once https://github.com/rucio/rucio/pull/5747 is merged. The changes @ivmfnal did in the closed PR are worth porting to the new code.

The first commit is extremely difficult to review. It mixes new functionality; with code move; with refactoring; and code formatter visual changes.

Should I re-assign this one to @cserf ?

this PR is based on top of #5915 . Only the last commit is the actual change.

Code formatting changes must not be submitted in the same commit as functional changes. This PR is impossible to review. I'm forced to ask you to update your PR to...