Brian O'Rourke

@mahmoudawadeen a few questions for you: 1. Is this still an issue for you, or did you come to some sort of resolution? 2. How frequently does this happen? 3....

I think [this comment](https://github.com/moove-it/sidekiq-scheduler/issues/320#issuecomment-762676536) handles this use case (i.e. you can just run `Sidekiq.reload_schedule!` rather than serializing back and forth to YAML).

@rvyas you might want to take a look at [sidekiq-uniq](https://github.com/meedan/sidekiq-uniq)

@rvyas I don't use that one myself, but I know others have had similar issues. Sidekiq Enterprise has [built-in support for unique jobs](https://github.com/mperham/sidekiq/wiki/Ent-Unique-Jobs), which would probably be your best choice...

@marcelolx this seems like a tricky thing for sidekiq-scheduler to handle well, as it's still always possible for the app to enqueue jobs externally from the scheduler.

I think there may be 3 issues here: A) Sidekiq-scheduler can enqueue a job when it's already running B) Sidekiq can enqueue a job when it's already running C) Sidekiq...

@beanieboi would you mind confirming if this is still an issue and/or what ended up happening?

Sounds good. If you'd like a patch from me for either implementation just ping me when you're ready and let me know which branch to target.

@johnnyshields the trace you shared shows three calls to `read` and `format_reply`. That wouldn't cause a `SystemStackError`. Some limited recursion is normal when the data being operated on is changing...

#679 describes the same issue - glad you're looking into a fix, but note that earlier efforts to resolve this ran into backwards-compatibility issues.