OK, I guess I'm gonna score some originality points here, but here goes:
After a long time with delayed_job, a project I'm involved in is experimenting with alternate queue adapters, namely the que
Since it has no explicit handling for long running tasks, I have wrapped execution in a timeout block similarly to what delayed_job does.
However, we are getting a lot of Prepared Statement already exists errors.
This might have something to do with the query being interrupted via the timeout and the postgres activerecord adapter not having the chance to advance the prepared statement counter, leading to following statements trying to use an already reserved identifier.
This is speculation at this point, however I would like to ask if delayed_job has any handling for gracefully recovering from interrupted queries, since it doesn't seem to suffer from the same issue.
Thanks in advance,