Re: A basic Shepherd bug?

[Felix Lechner]

Hi Attila,

On Fri, Mar 01 2024, Attila Lendvai wrote:

> you `guix system reconfigure` into a new shepherd version, and after
> that the currently running shepherd init process went 100% CPU,
> i.e. it was busy looping in one thread?

Yes, I used 'guix deploy.' I did so several times before noticing the
unusual load.

I don't usually reboot when changing services, but have to when adding
new ones.  I'm so lazy about it, sometimes it seems we should keep track
of kernel and Shepherd versions separately from other features of the
system. In effect, we would have two types of "system generations,"
namely:

  A. Kernel/PID 1 generation
  B. Features generation (not using the word "services" here)

> a quick and dirty solution

It was just a thought. Ideally, it would appear in a bug tracker so
someone can think about it with the benefit of hindsight later.

> interesting thoughts on migration and staged computation

I just wished for an alert when the code deployed as PID 1 had changed,
but no worries, please. There is so much to do.

Thanks for your (and Ludo's) hard work on the Shepherd. I love it!

Kind regards
Felix