Some points:
- nix-collect-garbage will collect garbage of the builtins /nix/store
path, right?
- you're case 2 (recovery) could be replaced by introducing a salt
such as passing a dummy var to a derivation everything depends on.
I agree that there are use cases which this doesn't catch
On 10/30/2011 11:19 AM, Peter Simons wrote:
Author: simons
Date: Sun Oct 30 15:19:58 2011
New Revision: 30127
URL: https://nixos.org/websvn/nix/?rev=30127sc=1
Log:
Reverting revisions 30103-30106: always set
nixpkgs.config.{state,store}Dir, etc.
After the change from revision 30103,
Hi Shea,
In the future, can you please bring up an issue like this on the
mailing list before just reverting another developer's work?
you are right, I should have just posted to the list instead of
reverting the commits. I was too impatient, and I overestimated the
severity of the problem.
What happened from my perspective is that I ran nixos-rebuild, and then
left the computer for something else. When I came back after a while,
the system had become nearly unusable. [..]
We all agree on that being a worse case.
So what's the true fix? Why can processes use 30GB of swap on your
Yes, indeed. I wonder whether it would be possible to configure useful
ulimits in the nix-daemon upstart job?
Yes - we should talk about what would be a sensible limit.
At least 512MB (for kernel builds or such) are required. So I'd vote for
something between 2 and 4GB - because there are some
Hi Marc,
Yes - we should talk about what would be a sensible limit. At least
512MB (for kernel builds or such) are required. So I'd vote for
something between 2 and 4GB - because there are some builds running
kvm and building nixos inside that.
personally, I would like to configure that
One could argue whether some limit should be enabled by default or not?
Yes - of course it should be an option - as always :)
I'd wait two days. Maybe people have reply having other requirements.
I only did a very rough estimation which should serve most cases I can
think of.
And in all other