2 years ago
{ kind: IOError(Os { code: 11, kind: WouldBlock, message: "Resource temporarily unavailable" }) }
I have a dagster instance deployed on Railway and using multithreading in Python. I also have dbt assets materializing and I am running out of resources, leading to the deployed instance crashing. I'm wondering if this is due to a resource limit on our Pro Railway plan? I've currently got max 8 workers for the multithreading job and 10 concurrent asset materializations allowed.
24 Replies
2 years ago
can you try lowering the workers and jobs?
We fixed this by restarting the container every 4 hours. Our software works by starting threads and processes in the container and we were hitting NPROC_LIMIT on the container. Raising the limit was not possible.
RLIMIT_NPROC=(soft:99999 hard:99999)
We suspect that this was unlimited before and now set to 100k
2 years ago
Where are you trying to migrate it to?
2 years ago
May I ask why you are trying to revert the Postgres version?
2 years ago
latest is 16 for our images
2 years ago
its possible the target region does not have the available space for that migration
2 years ago
Yes you would want to wait until Metal Volumes are GA for Pro users, currently it is only GA for Hobby users
2 years ago
hello, it is possible we just fixed this, could you try again?
2 years ago
thank you for trying
hello there, we are continuing to get this error
alembic.util.exc.CommandError: Version table 'alembic_version' has more than one head present; please use get_current_heads()
2 years ago
I'm really sorry but I'm not too sure how much we can help you with that as it's unlikely to be related to the Railway platform