Backfill: Missing required config entry

[the4thamigo-uk]

Hi I have a job running two partitioned assets A and B such that B depends on A. I have some partitions that have failed with a transitory error. When I try to backfill the failed partition using the UI, I get the following error :

Missing required config entry "ops" at the root. Sample config for missing entry ...

However, if I go the the run that failed and hit the Re-execute button, the run executes.

How do I fix this so I can successfully reprocess the partition using a backfill?

[sryza]

Hi @the4thamigo-uk - presumably you're hitting this error because one of your assets requires config? Are you backfilling via the job page, and does your job use PartitionedConfig?

[the4thamigo-uk]

Right, I think I understand. My sensor was setting the config for the job in order to pass the name of a file that it had detected in external s3 bucket. Obviously when I ran a backfill the config wasn't set. I have restructured the code so it doesn't use config in this way, although it does require me to have logic to rediscover the file again in the asset.

[hedeesaa]

I'm facing the same issue. How can we pass config to a backfilling job?
I checked the LaunchPartitionBackfill graphql query, it doesn't have accept run_config.