Rebuild is caching the flake

Development Deployments
1

I have infected a host via ssh and now have a bare NixOS server. I then apply my flake like this:

nixos-rebuild switch --flake git+ssh://git@github.com/tcurdt/nixcfg.git#utm-arm

and get the system I want. :tada:

But when I make a change to the repo and re-run the same command nothing happens.
I seems like the content for that repo is cached. Why is that? Is there a way around that?

In the long run it’s probably better to look into NixOps (or something) but I thought I start to learn the land like this first.

2

That’s the tarball-ttl nix setting, which can be set in /etc/nix/nix.conf, ~/.config/nix/nix.conf, or on the cmdline with --tarball-ttl. From man nix.conf:

The number of seconds a downloaded tarball is considered fresh. If the cached tarball is stale, Nix will check whether it is still up to date using the ETag header. Nix will download a new version if the ETag header is unsupported, or the cached ETag doesn’t match.

Setting the TTL to 0 forces Nix to always check if the tarball is up to date.

Nix caches tarballs in $XDG_CACHE_HOME/nix/tarballs.

Files fetched via NIX_PATH, fetchGit, fetchMercurial, fetchTarball, and fetchurl respect this TTL.

Default: 3600

1 Like
3

I knew there must be a way. Thanks!

Unfortunately at least nixos-rebuild switch does not seem to have a --tarball-ttl command line parameter. I guess I could set the ttl in my flake though.

4

Oh, yea, anything that can be set in nix.conf can also be set on the CLI like --option tarball-ttl 0. The --tarball-ttl argument is actually just shorthand for --option tarball-ttl, and not all of the CLI tools support this shorthand for every option.

1 Like
5

Hm. Are you sure?
Because I cannot find anything in the man page.
And trying various variations neither did work.

nixos-rebuild --option "tarball-ttl=0" switch --flake git+ssh://git@github.com/tcurdt/nixcfg.git#utm-arm
nixos-rebuild switch --flake git+ssh://git@github.com/tcurdt/nixcfg.git#utm-arm --option "tarball-ttl 0"
6

yea, I’m fairly sure. And it would be --option tarball-ttl 0, not --option tarball-ttl=0. I’m not sure why that wouldn’t be working. Unless maybe --flake doesn’t respect tarball-ttl? But I’m fairly sure it does. So I’m surprised it isn’t working for you.

7

Ah, Nix likes to cache things. I’ve solved this in the past by asking Nix to fetch the latest source for a flake uri:

nix flake prefetch github:my/repo

Then you can run whatever commands you need to that rely on that flake.

1 Like
8

Hm. I tried both versions (“tarball-ttl=0” and “tarball-ttl 0”) without success.
Also not sure which position to add the “–option” at - or whether that matters.

9

@tcurdt I ran into the same problem while developing my flake on one machine and continually doing nixos-rebuild switch on another. It seems there is an (undocumented, or at least not present in the man-page) option called --refresh, which seems to ensure the flake is refreshed from Github, even if the cache is still considered fresh. Hope it helps.

1 Like
10

--option tarball-ttl 0 (without the quotes) should work

1 Like