Question: prevent duplicate jobs

[sbeckeriv]

Dearest Maintainer,

I was wondering if there was a way to prevent duplicate jobs where the payload might be the same. In a hand made system we used a hash to generate an optional number. Each job provided its own value. If it was present we had a partial index on the hash and the channel name that prevented the inserting of the duplicate.

I have tried using the ordering flag for this but i have found that one job dies it all dies.

please let me know any thoughts.
Becker

[Diggsey]

sqlx does not retain jobs once they are complete, so any approach to deduplication is inherently racey.

I would suggest keeping track of already-spawned jobs in your own table and check that table in the same transaction that you are going to spawn the job. That way you can simply skip spawning if the job already exists.