https://b.tuxes.uk/three-years-of-ephemeral-nixos.html
HN: Three Years of Ephemeral NixOS: My Experience Resetting Root on Every Boot | Hacker News
Lobsters: Three Years of Ephemeral NixOS: My Experience Resetting Root on Every Boot | Lobsters
https://b.tuxes.uk/three-years-of-ephemeral-nixos.html
HN: Three Years of Ephemeral NixOS: My Experience Resetting Root on Every Boot | Hacker News
Lobsters: Three Years of Ephemeral NixOS: My Experience Resetting Root on Every Boot | Lobsters
I’m quite surprised to hear that bind mounts cause significant IO overhead. I wish this would have been quantified and measured.
I suppose that article’s remark is in reference to Significant Performance Impact · Issue #248 · nix-community/impermanence · GitHub, which has some benchmarks.
FWIW FUSE is known to be slow, but that’s a different issue.