This setting enables the database to be written to the working directory each time the database interface has an interaction. This basically ensures that a relatively current database is written if a job fails in a way that prevents going through interactError. There is potential overhead in the operation due to the file I/O. But I checked with a large model, where the runtime difference would be greatest, and it didn't have any noticeable impact on runtime.
What is the change?
syncDbAfterWrite
now defaults toTrue
.Why is the change being made?
This setting enables the database to be written to the working directory each time the database interface has an interaction. This basically ensures that a relatively current database is written if a job fails in a way that prevents going through
interactError
. There is potential overhead in the operation due to the file I/O. But I checked with a large model, where the runtime difference would be greatest, and it didn't have any noticeable impact on runtime.Checklist
doc
folder.pyproject.toml
.