My repo.py is structured like this,
# announcements
c
My repo.py is structured like this,
m
@sashank @prha
s
And what does your
workspace.yaml
look like?
c
Copy code
load_from:
  - python_file: repository.py
Just that, it seemed to be a problem with the migration. I havent managed to execute the migration correctly, but it works when I used a new Postgres DB
s
Is this similar to what you saw https://github.com/dagster-io/dagster/issues/2601?
cc @alex
c
Hm, not quite, i had this error
Copy code
sqlalchemy.exc.ProgrammingError: (psycopg2.errors.UndefinedTable) relation "runs" does not exist

[SQL: ALTER TABLE runs ADD COLUMN snapshot_id VARCHAR(255)]
Running the migration with this command
Copy code
$ PYTHONPATH=$PYTHONPATH:${PWD}/app DAGSTER_HOME=~/dagster/dagster_home dagster instance migrate
p
In your
dagster.yaml
, is your instance configured to use the same postgres DB for runs, event_log, and schedule storage or different DBs? Has that changed recently?
There’s an issue in one of our migration scripts, but just want to make sure the fix will unblock you…
c
In the configuration on which I tried the migration the dbs are different. In production we have a single RDS instance.
p
Okay, perfect… we’ll get a fix out for this today
c
Thanks! I'm still confused on whether the bug happens when you use multiple or single dbs?
p
The bug happens when you use multiple DBs… we were missing a check to make sure that a migration only happens on the runs storage
c
oh I see
p
This should be resolved in
0.8.2
. Lemme know if you are still seeing issues running
dagster instance migrate
c
Thanks, will try