Details
-
Type:
RFC
-
Status: Implemented
-
Resolution: Done
-
Component/s: DM
-
Labels:None
Description
RTN-013 describes a path for PanDA in the near term. This RFC is to publicise that fact and allow for comment on panda as a longer term solution.
From RTN-013:
We need a workflow system and production tools which can process DESC DC2 for DP0.2.
For PREOPS we should focus on PanDA in the near term This is separate from the construction side’s HSC reprocessing at NCSA for development needs. The construction team at NCSA can continue to use HTCondor-based BPS for the biweekly HSC reprocessing at NCSA.
My understanding for this evaluation is that:
I have heard talk that PanDA is a bit of a pain to administer which is why I want to be sure that Brookhaven are committing to do that support for the lifetime of the survey.
I'm happy for us to commit to PanDA for DP0.2 and re-evaluate. I am confident that other solutions would work as a back up but hope that the PanDA "added value" is compelling. The Oracle situation worries me in the long term but in some sense that is Richard Dubois's problem if SLAC are willing to support Oracle and also aren't charging the project for it.