Closed belforte closed 11 months ago
I did a quick test on CMSSW_9 on lxplus7 (python2) and it is OK. Once have test in Jenkins will run all archs and of course I need to make sure I did not break anything !
Jenkins results:
Jenkins results:
Jenkins results:
thanks a lot @novicecpp I have a question for you now: would it then be better to also move HTTPRequests to ClientUtilities ? Since it is not used by CRABRest only.
I merged my branch with master dmwm repo in test-setdatasetfilesstatus branch tagged as testSetDsetFilesStatus.1 and ran Jenkins CV to check for regression
All OK !
@novicecpp I captured your comments before in this list of actions. I will add more as needed once you answer my questions. Feel free to check and comment in there. Thanks
I have a question for you now: would it then be better to also move HTTPRequests to ClientUtilities ? Since it is not used by CRABRest only.
Maybe move HTTPRequests to utils and move DBSREST to new file same level as CrabRestInterface.py?
Jenkins results:
Jenkins results:
Jenkins results:
Jenkins results:
Jenkins results:
Jenkins results:
Hope I addressed all coments (only the uri
vs. api
confusion is left, but I think it is too much work). Will run Jenkins again.
New code tagged as testSetDsetFilesStatus.2
Jenkinks OK
Jenkins results:
code with last changes is now tagged as testSetDsetFilesStatus.2 in the ad-hoc test branch
Jenkins results:
ooopppss. I had added content-type in the wrong place. But of course CRAB Server REST happily ignores it and everything kept working, but better not confuse the reader.
If there are no more objection I will do a BIG SQUASH and merge.
Jenkins results:
thanks @mapellidario and @novicecpp for review and advice. I will merge. And we can discuss separately if/how to add --recursive. Maybe wait for someone to ask for it :grin: ?
We still have time to decide if to keep old semantic (Invalidating a dataset will automatically invalidate all the files in the dataset) or leave it to user with two actions as I indicated above. It seems a soft decisions, with arguments in pro/con either choice.
Since in current implementation crab setdatasetstatus
does not change status of individual files, I will print a note about this when the command is executed !
Thanks Stefano for the detailed description of dataset and file validity in DBS. Sorry for the dumb questions!
Do we already have an entry in the FAQs about DBS dataset and file status? Since we are allowing users to change it, we should also provide clear instructions about what those are and what the commands we provide do (or plan to do).
Dario,your questions were not dumb. The problem is that there is no real rule on how to use those status flags, DBS allows to set, then everybody can do as they see fit. I think Production changed their practice over time.
As user documentation this is what we have: https://twiki.cern.ch/twiki/bin/view/CMSPublic/CRAB3FAQ#Can_I_delete_a_dataset_I_publish it clearly needs a rewriting ! That is my next step.
Jenkins results:
Jenkins results:
I think this is OK now. Aside from