Martin Barisits

Let's keep this as `hostname` only (no `host`) to be consistent with the previous client. (To me, hostname and host are not the same, the one is a human readable...

Hi Alex, Yes this is a good idea, but needs to be made configurable as Mario said.

I am not sure where the issue lies yet. The problem is, as Mihai pointed out, that the submission format for multi-source jobs is bad in the first place. Because...

So in principle we do provide "corrected" data on this and Rucio sends out a message based on the src_url FTS chose: https://github.com/rucio/rucio/blob/5417d0954f49817d89b058462fbec4929f238656/lib/rucio/transfertool/fts3.py#L626-L641 But for this to work, that data...

@mpatrascoiu @murrayc3 one thing we could do is to extend the `metadata` (file_metadata) dictionary for the special case of a multi-source job. Currently it is build like this: ``` 'metadata':...

Coming back to this. > I am going to ask the CMS experiment to stop writing incorrect RSE values in the opaque FTS file-metadata field. I personally believe no value...

Consider it as sponsored by @labkode and me 😄 I do consider this as something of general scope to solve in Rucio/FTS, it's not experiment specific. But we need to...

Not yet, let's give @labkode some time to investigate. I don't think this is very complex, but Hugo should comment when his schedule allows to look into the issue. And...

I think we should perhaps follow both approaches. a) Might work for ATLAS and CMS but we might impact the monitoring of other communities, without some type of enrichment procedure,...

I think this makes sense. When @jamesp-epcc and I originally discussed this, we were thinking to make it as easy as possible. But perhaps we need to refine this.