I’m a newbie to dagster…so I’m running dagster 0.1...
# ask-community
d
I’m a newbie to dagster…so I’m running dagster 0.14.5 and am seeing this error: Sensor openlineage_sensor was started from a location repos.py that can no longer be found in the workspace, or has metadata that has changed since the sensor was started. You can turn off this sensor in the Dagit UI from the Status tab. Here is repo.py
from dagster import repository
from openlineage.dagster.sensor import openlineage_sensor
from hello_cereal import hello_cereal_job
@repository
_def_ hello_cereal_repository():
openlineage_sensor_def = openlineage_sensor(
_minimum_interval_seconds_=60,
_record_filter_limit_=60,
)
return [hello_cereal_job, openlineage_sensor_def]
Thoughts?
d
Hi Dan - is upgrading to the latest version of dagster an option? (0.14.9). Some changes went in recently that should make it a lot harder to run into that specific error
d
I tried that, but the openlineage-dagster plugin isn’t happy. So I reverted.
d
Are there any more details you can share about what didn't work when you upgraded?
d
If I run “pip install dagster==0.14.9”, I get this:
ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
openlineage-dagster 0.8.0 requires dagster<=0.14.5,>=0.13.8, but you have dagster 0.14.9 which is incompatible.
dagster-graphql 0.14.5 requires dagster==0.14.5, but you have dagster 0.14.9 which is incompatible.
d
Got it - i'd be surprised if there was actually a breaking change between 0.14.5 and 0.14.9 that would affect openlineage-dagster, we can probably work with them to make that pin less strict. Despite pip complaining, does dagster possibly still work if you run "pip install dagster==0.14.9 dagster-graphql==0.14.9"
d
pip stopped complaining, but I’m still seeing my original error from the daemon. More info: dagit recognizes the sensor just fine, but dagster-daemon is being contrary.
d
you might need to run 'dagster instance migrate' once after upgrading to make it go away
the other thing you'll want to check is that you're starting dagit and your daemon with the same arguments
and from the same folder
d
1. all shells have DAGSTER_HOME set the same. 2. I’m starting dagit and dagster-daemon with no arguments 3. They are in the same folder
d
Does the problem persist if you stop the sensor in dagit and then restart it?
And this is after running
dagster instance migrate
?
(and restarting the daemon)
d
Problem persists: Yes After run migrate: Restarted daemon and dagit
d
And if you run 'pip freeze | grep dagster' it shows that 0.14.9 is installed?
Ah, and dagit as well should be at 0.14.9 - I should have mentioned that too
although i wouldn't expect that to causes the problem
d
Yes and Yes. I followed the directions here to set up the sensor. I’m not a dagster expert, so I followed their directions.
d
Is it possible to share or DM the schedules.db file in your DAGSTER_HOME folder?
I'm starting dagit and dagster-daemon with no arguments
Does this mean that you've set up a
workspace.yaml
file in that same folder that you're running dagit from? is it possible to share the contents of that file?
d
Here is schedules.db and I’ve got an empty workspace.yaml. The docs weren’t really clear about what had to be here.
d
Got it - When you originally turned the sensor on, did you launch dagit with different arguments? Possibly
dagit -f repos.py
?
d
I’m an idiot…Here is my worksapce.yaml:
load_from:
- python_file: repos.py
I looked in the wrong file..sorry
d
Ah no worries, that makes more sense
One other question - are dagit and the daemon running in the same python environment? i.e. would both of them think that the python executable was /Users/dan.mahoney/Projects/dagster-test/bin/python3?
d
No
d
Ah, that will do it. We're working on eliminating that requirement that they need to be in the same environment (I actually thought we had already eliminated it, but it looks like there are a couple of small tweaks left to get it over the finish line)
d
I will play and get back to you.
Thanks for the help.
d
No problem, sorry for the rough edge
d
Just wanted to update. The environment was the culprit. Now on to the next problem.
d
Got it - thanks for flagging that. We just landed a fix that will go out in the release tomorrow that should make it work when they're in different environments
d
Cool…no more tripping up newbies.