https://dagster.io/ logo
#ask-community
Title
# ask-community
b

Barry Sun

06/20/2022, 12:13 AM
Hi all, I'm trying to get to grips with Dagster Instance/Daemon. I have looked through the docs here (https://docs.dagster.io/deployment/dagster-daemon) and concluded that I need a dagster instance (
dagster.yaml
) file to define which daemons to run, which is located in the same folder as
workspace.yaml
. I am trying to run a Backfill, for this which daemon do I need, and how do I define it in my dagster instance file?
🤖 1
👍 1
All good! Turns out I need to add run_coordinator like so
Copy code
run_coordinator:
  module: dagster.core.run_coordinator
  class: QueuedRunCoordinator
and it covers schedules, sensors, backfills 🙂
Something that I did struggle with, was that my
workspace.yaml
file is located in folder that contains the
pipeline
directory, which is actually my desired working directory that contains my
repository.py
file. By adding the working directory to my workspace file, it doesn't actually shift the the working directory to
pipeline
which was messing with my io_manager. A work around was to add the working directory as an input to my io_manager. Hopefully this makes sense, I just want to check if this is intended and whether the work around is best practice?
d

daniel

06/20/2022, 2:02 PM
Hi Barry - I think the backfill daemon will always be set up as long as you're running the daemon process
1
setting the working_directory in your workspace will affect where it looks for imports when loading your code, but won't necessarily affect the base directory of your io manager
1
b

Barry Sun

06/20/2022, 11:15 PM
got it - thanks Daniel!