bug#54453: crash running 'guix pull'

[zimoun]

Hi,

On Mon, 21 Mar 2022 at 16:23, Peter Bhat Harkins <peter@push.cx> wrote:

> The VPS was created almost two years ago, and 'guix pull' was run many
> times, though not in the last few months.

Did you garbage collect “guix gc”?  For instance with the option ’-F 5G’
or this option ’-d 9m’.

Because if you pull again and again, the store is growing.  And
depending on the size your VPS, you can reach the limit of the available
space.


> Is there documentation anywhere of what system resources a guix system,
> or this particular command, requires? Is there any method for
> estimating? It sounds like you're suggesting the requirements have
> changed over time. Do running guix systems require more disk and RAM as
> they age, or does this undocumented requirement also apply to new
> systems?

No, I am not suggesting a change.  I am trying to identify why your
setup is failing and emit hypothesis.

> At this point I'm less concerned about this individual crash bug and
> more concerned that running any guix system is unsustainable.

You are drawing hard conclusions from some hypothesis when they are just
hypothesis.

Over this 2 years, how many times did you garbage collect?


Cheers,
simon