Using an http proxy, derivation failed

Hi,

I am trying to get Guix in a QubesOS/debian 11 template. QubesOS expose a proxy

in its templates, used mostly by apt and dnf.

For some reason all my attempts to modify the systemd unit were

un-successful so I simply exported it on the command line. However as

you can see in the following backtrace, that is not working as expected.

$ export https_proxy=http://localhost:8082/; guix pull

hint: Consider installing the `glibc-locales' package and defining `GUIX_LOCPATH', along these lines:

guix install glibc-locales

export GUIX_LOCPATH="$HOME/.guix-profile/lib/locale"

See the "Application Setup" section in the manual, for more info.

Updating channel 'guix' from Git repository at 'https://git.savannah.gnu.org/git/guix.git'...

Building from this channel:

guix https://git.savannah.gnu.org/git/guix.git 515406f

Computing Guix derivation for 'x86_64-linux'... \substitute: hint: Consider installing the `glibc-utf8-locales' or `glibc-locales' package and

substitute: defining `GUIX_LOCPATH', along these lines:

substitute:

substitute: guix install glibc-utf8-locales

substitute: export GUIX_LOCPATH="$HOME/.guix-profile/lib/locale"

substitute:

substitute: See the "Application Setup" section in the manual, for more info.

substitute:

hint: Consider installing the `glibc-utf8-locales' or `glibc-locales' package and

defining `GUIX_LOCPATH', along these lines:

guix install glibc-utf8-locales

export GUIX_LOCPATH="$HOME/.guix-profile/lib/locale"

See the "Application Setup" section in the manual, for more info.

guix substitute: error: host name lookup error: Name or service not known

Backtrace:

14 (primitive-load "/gnu/store/m8ipcivlnaqr4y9hpxiiqi2pn1wldcia-compute-guix-derivation")

In ice-9/eval.scm:

155:9 13 (_ _)

159:9 12 (_ #(#(#(#(#(#(#(#(#(#(#(#(#(#(#(#(#<directory (guile-u?> ?) ?) ?) ?) ?) ?) ?) ?) ?) ?) ?) ?) ?) ?) ?) ?))

In ice-9/boot-9.scm:

152:2 11 (with-fluid* _ _ _)

152:2 10 (with-fluid* _ _ _)

In ./guix/store.scm:

2129:24 9 (run-with-store #<store-connection 256.99 7400779ab140> #<procedure 740061d42d20 at ./guix/self.scm:12?> ?)

1966:8 8 (_ #<store-connection 256.99 7400779ab140>)

In ./guix/gexp.scm:

296:22 7 (_ #<store-connection 256.99 7400779ab140>)

1180:2 6 (_ #<store-connection 256.99 7400779ab140>)

1046:2 5 (_ #<store-connection 256.99 7400779ab140>)

892:4 4 (_ #<store-connection 256.99 7400779ab140>)

In ./guix/store.scm:

2014:12 3 (_ #<store-connection 256.99 7400779ab140>)

1406:5 2 (map/accumulate-builds #<store-connection 256.99 7400779ab140> #<procedure 740072c0d640 at ./guix/stor?> ?)

1421:15 1 (_ #<store-connection 256.99 7400779ab140> ("/gnu/store/gmi62pbnf0jfish26chd7pvfzs2rzlxa-guile-ssh-?" ?) ?)

1421:15 0 (loop #f)

./guix/store.scm:1421:15: In procedure loop:

ERROR:

1. &store-protocol-error:

message: "some substitutes for the outputs of derivation `/gnu/store/ak595vg1ij7vf3vxx3srk60dvwz6zxry-gnutls-3.7.2.drv' failed (usually happens due to networking issues); try `--fallback' to build derivation from source "

status: 1

guix pull: error: You found a bug: the program '/gnu/store/m8ipcivlnaqr4y9hpxiiqi2pn1wldcia-compute-guix-derivation'

failed to compute the derivation for Guix (version: "515406f20f0f2a69f8701ce3ece8121dbc4fb5dd"; system: "x86_64-linux";

host version: "4578ad78bee1dfd450962627ae66d3e15ebadad2"; pull-version: 1).

Please report the COMPLETE output above by email to <bug-guix@gnu.org>.

Hello,

dal-blazej--- via Bug reports for GNU Guix <bug-guix@gnu.org> writes:

As noted in the manual, the guix daemon itself (and not the commands

that communicate with it) needs to have the https_proxy environment

variable set. What went wrong while modifying the systemd unit?

Best,

--

Josselin Poiret

Hi debbugs,

The following mails was not CC to 53904@debbugs.gnu.org so let's correct

that :

Hi Josselin,

Josselin Poiret <dev@jpoiret.xyz> writes:

The systemd unit file from debian looks like that

(I only added the https_proxy, with localhost or 127.0.0.1, with or

without the leading http://):

#+begin_src conf :tangle "/etc/systemd/system/guix-daemon.service"

# This is a "service unit file" for the systemd init system to launch

# 'guix-daemon'. Drop it in /etc/systemd/system or similar to have

# 'guix-daemon' automatically started.

[Unit]

Description=Build daemon for GNU Guix

[Service]

ExecStart=/var/guix/profiles/per-user/root/current-guix/bin/guix-daemon --build-users-group=guixbuild

Environment='GUIX_LOCPATH=/var/guix/profiles/per-user/root/guix-profile/lib/locale' LC_ALL=en_US.utf8 'https_proxy=http://localhost:8082/'

RemainAfterExit=yes

StandardOutput=syslog

StandardError=syslog

# See https://lists.gnu.org/archive/html/guix-devel/2016-04/msg00608.html.

# Some package builds (for example, go@1.8.1) may require even more than

# 1024 tasks.

TasksMax=8192

[Install]

WantedBy=multi-user.target

#+end_src

After reloading the daemons and restarting the daemon :

guix pull: error: failed to connect to `/var/guix/daemon-socket/socket': Connection refused

By the way, I previously tested the proxy itself via wget and it

works. I also tried some others variations on the unit file without more success.

Hello,

dal-blazej@onenetbeyond.org writes:

I'm not very familiar with the Debian Guix package, but can you check

that /var/guix/profiles/per-user/root/current-guix/bin/guix-daemon does

indeed exist? If not, you may need to log-in as root, and run `guix

pull` there. Did the systemd unit ever work, and if not, how do you run

the Guix daemon?

Best,

--

Josselin Poiret

Thanks for your help Josselin.

I precise that I am able to run guix in a debian standaloneVM without proxy.

I feared to have messed up something in that template so I started again

from scratch.

- I modified /usr/lib/systemd/system/guix-daemon.service and

/usr/lib/systemd/system/guix-publish.service to use

https_proxy=http:///127.0.0.1:8082

- /var/guix/profiles/per-user/root/current-guix doesn't exist.

- guix pull (as user or root) :

#+begin_exemple

Updating channel 'guix' from Git repository at 'https://git.savannah.gnu.org/git/guix.git'...

guix pull: error: Git error: failed to resolve address for

git.savannah.gnu.org: Temporary failure in name resolution

#+end_exemple

I am still able to use my proxy with wget.

- guix-daemon.service is now running without error.

- guix-publish.service throw an error (but I guess it is another issue):

#+begin_exemple

? guix-publish.service - Publish the GNU Guix store

Loaded: loaded (/lib/systemd/system/guix-publish.service; enabled; vendor preset: enabled)

Active: failed (Result: exit-code) since Thu 2022-02-17 16:44:00 CET; 7min ago

Process: 7397 ExecStart=/usr/bin/guix publish --user=nobody --port=8181 (code=exited, status=1/FAILURE)

Main PID: 7397 (code=exited, status=1/FAILURE)

CPU: 184ms

Feb 17 16:44:00 localhost systemd[1]: Started Publish the GNU Guix store.

Feb 17 16:44:00 localhost guix[7397]: guix publish: error: open-file: No such file or directory: "/etc/guix/signing-key.pub"

Feb 17 16:44:00 localhost systemd[1]: guix-publish.service: Main process exited, code=exited, status=1/FAILURE

Feb 17 16:44:00 localhost systemd[1]: guix-publish.service: Failed with result 'exit-code'.

#+end_exemple

Hi,

dal-blazej@onenetbeyond.org skribis:

[...]

Looks good to me.

This suggests that guix-daemon is not running. Could it be that it

failed to start? Does ‘journalctl -u guix-daemon.service’ contain any

clue?

BTW, according to the output you pasted earlier (in particular the noisy

locale warnings), it seems that your daemon needs an update. Check out

https://guix.gnu.org/manual/devel/en/html_node/Upgrading-Guix.html.

Thanks,

Ludo’.

Hi,

Ludovic Courtès <ludo@gnu.org> writes:

Yes It needs an update but it can't. That's the issue ;)

Sorry, I think my CC error has lead to some confusion.

My last mail stated I started from scratch (new virtual machine, new

guix installation) :

- with the https_proxy defined inside the systems service file.

- with the daemon running

So please, ignore the first errors.

In this case the error seems clearly that :

$ guix pull

use git, and _guix does not pass to git the https_proxy_.

Fortunately I can update by passing the https_proxy on the command:

$ https_proxy=https://127.0.0.1:8082guix pull

To make it clear, both the variable in the systemd service file and on

the command line are necessary to successfully run guix pull.

But as you stated it shouldn't, it still looks like a bug.