Hello everyone ! We deployed Dagster stack using Helm, and everything seems to work fine But once a job running all our assets is finished, we cannot re-materialize an asset depending on another one because it's gonna run the error

[Errno 2] No such file or directory: '/opt/dagster/dagster_home/storage/previousAsset'

with

previousAsset

the asset that's needed to rematerialize our asset We don't have this issue if we run it with

dagster dev

locally Any idea what's happening here ? Is the docker flushing previous data the way it's deployed ?

What IO manager are you using? If you're not using a IO manager which writes materializations to remote storage (S3, Cloud Storage, DB, etc.), then it's likely that the materialization information is being stored locally on the pod that was spun up to run your materialization job. That pod is spun down when the run finishes, at which point you'll lose the materialization information that was written to disk.

Perfectly clear, must be it then ! Thank you for your quick answer Zach!