Also, if I use an IOManager (say S3), then can I re-run specific solids? or just a part of the pipeline? If I do that, the intermediate data will be fetched from S3 right?
y
yuhan
08/04/2021, 7:41 PM
yes - as long as it’s a persistent IO Manager, dagster machinery would handle the logic for loading from previously persisted intermediate data.
m
madhurt
08/04/2021, 7:44 PM
so the
s3
IOManager is a persistent one right?
I am asking because there will be no data that will be stored locally. All intermediate outputs will be in S3 through this IOManager. Now if a pipeline has solids A --> B --> C and if B fails
and now I need to re-run B onwards i.e. B and C. So B should get A’s data through S3 right? Is that expected with the