Closed adamhamlin closed 1 year ago
Hmmm, 1 of the travis jobs passed, but the other 2 failed (on different tests, tho both singleton-related). Locally, I get no failures. Any ideas @timgit? resolved
I tried something similar in the past, read here why there might be „race conditions“ when 2 workers arrive at the same time. https://github.com/timgit/pg-boss/issues/274#issuecomment-932771001
I tried something similar in the past, read here why there might be „race conditions“ when 2 workers arrive at the same time. #274 (comment)
I could be missing something, but that comment talks about 2 concurrent workers selecting two different queued jobs, but for singleton queue there can only be 1 queued job to begin with.
right, than this might work 😄 I have not worked with those queues yet. I just recalled my old issue where I tried to do something similar (1 concurrent job per queue across multiple workers) - so I thought I just throw it in 😉
right, than this might work 😄 I have not worked with those queues yet. I just recalled my old issue where I tried to do something similar (1 concurrent job per queue across multiple workers) - so I thought I just throw it in 😉
I appreciate it @phips28! Thank you
@timgit just confirming you're not waiting on me for anything
@timgit sorry for the spam, just hoping to move this toward resolution
See relevant issue: https://github.com/timgit/pg-boss/issues/367
This PR would change the behavior of "singleton queue" jobs (i.e.,
singletonKey
+useSingletonQueue=true
) as follows:My assumption is that this is a bug--or at least not intentional or desired behavior--and so it does not warrant a breaking change or the need for configurability.