Matt Reiferson

Yep, it also doesn't present a confirmation dialog.

@ploxiln catching up on this... Doesn't this effectively _reduce_ the likelihood of messages being written to disk when `mem-queue-size < 32`, which is the opposite of what most users intend?...

sorry I'm late on this... I think I follow what's happening now, but I'm still not sold. In the event consumers are slow, this creates undesirable back-pressure when you'd rather...

We use `os.Link` to ensure atomicity for filesystems that support it. My recommendation here would be to have something else responsible for uploading to S3 and cleaning up the `--output-dir`.

I don't think so — we would have to rethink the behavior of other parts of the code given the lack of atomicity. However, we could consider at least documenting...

thoughts @jehiah ?

> It feels like throwing away messages on disk would be more natural based on the count of messages on disk. Would that target the same need here, or is...

@jehiah these are interesting ideas - I think there was some IRC discussion recently debating some of this separation. I think the _real_ tricky part is that all of this...

@stephensearles thanks again for putting this proposal together, apologies for the delay in responding. I've read through it and there are a lot of solid ideas. I think my biggest...

> The "backup" messages propagated would need to be stored apart from regular messages being published through an ansqd node. For the FIN to happen well, they'd need to be...