bug#49205: [services]: nginx doesn't restart after guix system reconfigure

[Ludovic Courtès]

Hi,

Tobias Geerinckx-Rice <me@tobias.gr> skribis:

> Raghav Gururajan via Bug reports for GNU Guix 写道：
>> Throw to key `%exception' with args `("#<&invoke-error program:
>> \"/gnu/store/g6nhdp7i1sr51anq86hd6adrsm5ha857-nginx-1.21.0/sbin/nginx\"
>> arguments: (\"-c\"
>> \"/gnu/store/2igpjf27rm92np6jn6hfrxzx5m55a468-nginx.conf\" \"-p\"
>> \"/var/run/nginx\") exit-status: 1 term-signal: #f stop-signal:
>> #f>")'.
>
> The shortest answer is to run this command manually and see what it
> prints out.  So
>
>  $ sudo /gnu/store/g6n…/sbin/nginx -c /gnu/store/… -p   /var/run/nginx
>
> in a terminal.

These days, with the Shepherd 0.9.3, the stdout/stderr or nginx is
captured in /var/log/messages when it starts.  That should make it
easier to debug startup failures.

If you don’t have more info on this bug Raghav (it’s been some time…), I
suggest we close it.

> I saw you mention nginx and certbot in #guix.  It's unfortunately
> common to end up in a catch-22 with Guix's nginx and certbot services
> when using HTTP authentication, where nginx needs to be running to
> receive your first certificate but refuses to start because of missing
> certificates…

I wonder how we could address that.

Ludo’.