-
Notifications
You must be signed in to change notification settings - Fork 43
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
stuck on old nixos generation due to nix version incompatibility #363
Comments
I managed to get around it with: # Uses the flake registry for Nix repo
# Requires building Nix HEAD from source (but can prob specify a different branch)
# HEAD is not guaranteed to work properly
alias nix='nix run nix --'
nix build /etc/nixos#nixosConfigurations."ryosuke".config.system.build.nixos-rebuild
sudo ./result/bin/nixos-rebuild boot [Edited from my original comment for accuracy] |
This is probably relevant data when discussing dynamic vs static linking of |
Correction: it is not so simple. I am also updating another machine and running into the same issue, even when running that Considering this seems to me like a blocker without an obvious solution, can the error message provide some further advice about how to proceed? I am guessing that anyone who is using Standard for their system flake configuration would run into this trap if they had an outdated system. Update: This other system (called ✖1 nix profile install nixpkgs#nix_2_13
error: flake 'flake:nixpkgs' does not provide attribute 'packages.x86_64-linux.nix_2_13', 'legacyPackages.x86_64-linux.nix_2_13' or 'nix_2_13'
Did you mean one of nix_2_3, nix_2_4, nix_2_5 or nix_2_6? |
I updated the original example to include my current workaround using |
There's no way to tell |
@blaggacao From what I can tell, there's nothing so obvious or direct, which certainly supports your description of Regardless, even if I could specify a Since this is feeling more like a log of my journey (and I'm done with updating old systems now), I could close this issue, though I still feel as though there is something strange in the way that Standard blocks In the end, it effectively leads me towards thinking I should avoid Standard for anything but the most up-to-date use cases with extensive CI updates automation etc. (it's hard to keep up sometimes) and not so much for personal NixOS machines that could sit in a closet for a while, subject to decay (this one laptop I'm working on had its clock stuck back in October 2022 cos of some WIP system generation breaking DNS...)). I guess that's kinda the point of Hive but in the end I just want the project structure to work reliably even if it's not updated in a long time (and, really, I don't think a year is that long of a time). |
If understanding is accurate, there are two failure modes:
Maybe both problems could be tackled by ignoring system-available Wouldn't that bring in the right version of The system, by pinning the Nix version to reproduce itself, effectively forbids language innovation. I think Nixlang should be mentally considered exogenous to the system. Thus, it might have even been a better design choice for |
i'm trying to update a nixos machine that's been sitting in my closet for a year. however, any time i try to run
nixos-rebuild
i get the following:and yet:
this has become increasingly frustrating as i have tried many workarounds, and all i want to do is update, but i seem to be stuck in some kind of catch-22 limbo situation
i'm not even sure why this error only shows up with
nixos-rebuild
-- i am not in a devshelli'm hesitant to run the nixos activation script manually because it's been so long
The text was updated successfully, but these errors were encountered: