You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I'm sorry in advance if this is not the right RFC to reference!
I opened an issue in flux-framework/flux-core #3136 that discussed some of the gaps when trying to create job usage reports. One of the fields needed is a bank field, a field that stores what bank the user charged their job against (filtering job records by bank was one of the features discussed with Ryan Day). @grondo left this feedback in that thread:
We'll need to determine a spot where the group/account gets set in jobspec. Then you can pull it from there, or it may be important enough to add to job-info and perhaps broken out as a separate key in job-archive?
We may want to open an issue on this in the rfc repo.
Does it make sense to place a bank field in jobspec? I assume that a bank would be an option passed in by the user when submitting a job, and it defaults to some bank if nothing is passed in:
$ flux mini submit --bank=this_specific_bank -n 1 hostname
From a flux-accounting perspective, I tried to think of ways to pull the same information directly from the flux-accounting database without needing to include it in jobspec. I first thought that a lookup could be performed with the username passed in to see which bank the user belongs in, but one issue that comes to mind is the possibility that a user belongs to multiple banks... so, would there instead to be a lookup in the flux-accounting database performed immediately after a job is submitted to validate which bank was specified?
The text was updated successfully, but these errors were encountered:
Does it make sense to place a bank field in jobspec? I assume that a bank would be an option passed in by the user when submitting a job, and it defaults to some bank if nothing is passed in:
👍 Makes sense to me. Maybe under attributes.system.bank?
At some point, we can take a deeper look at the scheduler-specific data schema like queue and bank. But for now, what @SteVwonder suggested makes sense to me.
I'm sorry in advance if this is not the right RFC to reference!
I opened an issue in flux-framework/flux-core #3136 that discussed some of the gaps when trying to create job usage reports. One of the fields needed is a
bank
field, a field that stores what bank the user charged their job against (filtering job records by bank was one of the features discussed with Ryan Day). @grondo left this feedback in that thread:Does it make sense to place a
bank
field in jobspec? I assume that abank
would be an option passed in by the user when submitting a job, and it defaults to some bank if nothing is passed in:From a flux-accounting perspective, I tried to think of ways to pull the same information directly from the flux-accounting database without needing to include it in jobspec. I first thought that a lookup could be performed with the username passed in to see which bank the user belongs in, but one issue that comes to mind is the possibility that a user belongs to multiple banks... so, would there instead to be a lookup in the flux-accounting database performed immediately after a job is submitted to validate which
bank
was specified?The text was updated successfully, but these errors were encountered: