Brendan (btrotter.delete@this.gmail.com) on January 13, 2020 7:21 am wrote:
> Consider a HTTP server - do you want to (e.g.) drop one connection or abort
> one request, or do you want to get killed and drop all current connections?

On Apache you can chose the behavior you prefer depending on the MPM you choose but generally having one process per connection will lead to more memory use, not less, even though recovering from a single process crash will have minimal impact.

> Consider a word processor - do you want to (e.g.) free memory from the "undo buffer" and retry
> (and if that doesn't work display a dialog box informing the user and save the current document
> and shut down gracefully); or do you want to get killed and lose all the user's unsaved work;
> or do you want kill X11 and screw up every single app that's currently running?

Detecting how close you are to an OOM situation is very, very hard. I know because I spent the last two years trying to figure out ways of doing it across multiple platforms. The only platform where it's easy to implement is Android because it has kernel-based notifications that are tied to the OOM killer so you know when it's going to kick in a few moments before it does.

> How about a database management engine with 8 GiB of cached data it can easily discard - would you
> want everything that depends on it to suffer "sudden database un-availability" for no reason?

A database needs to be crash-proof regardless of what type of crashes might happen to it. And besides with LMDB being available on pretty much any platform out there there's no excuse not to have crash-proof storage.