Empty Ledger Bug
In the gravelamps_background_study
in which many events have been submitted simultaneously (~200 PE runs for 100 events). After several chains were killed due to a 'conda endpoint not connected' issue which I presume to be the result of maintainence weirdness, I needed to reset the runs. The following procedure was followed:
-
condor_rm
all jobs - tried an
asimov manage submit
in the hopes that would automate the restart - when that did nothing quickly checked the result of
asimov monitor
which suggested the jobs were still running (now I did interrupt this to prevent it from doing all 200 productions) -
ledger.yml
was then empty
I suspect that it was probably interrupting asimov monitor
that did it which I take full responsibility for, but just to make you aware.