unofficial mirror of guix-patches@gnu.org 
 help / color / mirror / code / Atom feed
From: "Ludovic Courtès" <ludo@gnu.org>
To: "Jakob L. Kreuze" <zerodaysfordays@sdf.lonestar.org>
Cc: 36555@debbugs.gnu.org
Subject: [bug#36555] [PATCH 1/2] guix system: Add 'reconfigure' module.
Date: Sat, 13 Jul 2019 12:23:20 +0200	[thread overview]
Message-ID: <87y3129qsn.fsf@gnu.org> (raw)
In-Reply-To: <87ef30i9fl.fsf@sdf.lonestar.org> (Jakob L. Kreuze's message of "Mon, 08 Jul 2019 15:59:58 -0400")

Hello!

zerodaysfordays@sdf.lonestar.org (Jakob L. Kreuze) skribis:

> * guix/scripts/system/reconfigure.scm: New file.
> * Makefile.am (MODULES): Add it.
> * guix/scripts/system.scm (bootloader-installer-script): Export variable.
> * gnu/machine/ssh.scm (switch-to-system, upgrade-shepherd-services)
> (install-bootloader): Delete variable.
> * gnu/machine/ssh.scm (deploy-managed-host): Rewrite procedure.

[...]

> +  (define (run-switch-to-system machine)
> +    "Monadic procedure serializing the items in MACHINE necessary to build a
> +G-Expression with 'switch-to-system'."
> +    (mlet %store-monad ((script (switch-system-program (machine-system machine))))
> +        (machine-remote-eval machine #~(primitive-load #$script))))
> +
> +  (define (run-upgrade-shepherd-services machine)
> +    "Monadic procedure serializing the items in MACHINE necessary to build a
> +G-Expression with 'upgrade-shepherd-services'."
> +    (mlet* %store-monad ((target-services target-services)
> +                         (script (upgrade-services-program target-services)))
> +      (machine-remote-eval machine #~(primitive-load #$script))))

These would look nicer if ‘switch-system-program’ and
‘upgrade-services-program’ returns a <program-file> because you could
just write:

  (machine-remote-eval #~(primitive-load #$(switch-system-program …))
                       machine)

(I realize the order of arguments is reversed; to stick to what ‘eval’
does, I’d tend to put the ‘machine’ argument second—but that’s a
separate issue.  :-))

> +(define (switch-system-program os)
> +  "Return as a monadic value a derivation to build a scheme file that, upon
> +being evaluated, will create a new generation for SYSTEM-DERIVATION and
> +execute ACTIVATION-SCRIPT."
> +  (gexp->script
> +   "switch-to-system.scm"
> +   (with-extensions (list guile-gcrypt)
> +     (with-imported-modules (source-module-closure '((guix config)
> +                                                     (guix profiles)
> +                                                     (guix utils)))
> +       #~(begin
> +           (use-modules (guix config)
> +                        (guix profiles)
> +                        (guix utils))
> +
> +           (define %system-profile
> +             (string-append %state-directory "/profiles/system"))
> +
> +           (let* ((number (1+ (generation-number %system-profile)))
> +                  (generation (generation-file-name %system-profile number)))
> +             (switch-symlinks generation #$os)
> +             (switch-symlinks %system-profile generation)
> +             (setenv "GUIX_NEW_SYSTEM" #$os)
> +             (with-output-to-string
> +               (lambda ()
> +                 (primitive-load
> +                  #$(operating-system-activation-script os))))))))))

Can we remove ‘with-output-to-string’?  I’d rather see what’s going on.
:-)

If that’s too verbose, we can use ‘invoke/quiet’.

> +;; XXX: Currently, this does NOT attempt to restart running services. See
> +;; <https://issues.guix.info/issue/33508> for details.
> +(define (upgrade-services-program target-services)
> +  "Return as a monadic value a derivation to build a scheme file that, upon
> +being evaluated, will use TARGET-SERVICES, a list
> +of (shepherd-service-canonical-name, shepherd-service-file) pairs to determine
> +which services are obsolete and need to be unloaded, as well as which services
> +are new and need to be started."
> +  (gexp->script
> +   "upgrade-shepherd-services.scm"
> +   (with-imported-modules '((gnu services herd))
> +    #~(begin
> +        (use-modules (gnu services herd)
> +                     (srfi srfi-1))
> +
> +        (define running
> +          (filter live-service-running (current-services)))
> +
> +        (define (essential? service)
> +          ;; Return #t if SERVICE is essential and should not be unloaded
> +          ;; under any circumstance.
> +          (memq (first (live-service-provision service))
> +                '(root shepherd)))
> +
> +        (define (obsolete? service)
> +          ;; Return #t if SERVICE can be safely unloaded.
> +          (and (not (essential? service))
> +               (every (lambda (requirements)
> +                        (not (memq (first (live-service-provision service))
> +                                   requirements)))
> +                      (map live-service-requirement running))))
> +
> +        (define to-unload
> +          (filter obsolete?
> +                  (remove (lambda (service)
> +                            (memq (first (live-service-provision service))
> +                                  (map first '#$target-services)))
> +                          running)))
> +
> +        (define to-start
> +          (remove (lambda (service-pair)
> +                    (memq (first service-pair)
> +                          (map (compose first live-service-provision)
> +                               running)))
> +                  '#$target-services))
> +
> +        ;; Unload obsolete services.
> +        (for-each (lambda (service)
> +                    (false-if-exception
> +                     (unload-service service)))
> +                  to-unload)
> +
> +        ;; Load the service files for any new services and start them.
> +        (load-services/safe (map second to-start))
> +        (for-each start-service (map first to-start))))))

It seems that this sort-of inlines parts of ‘shepherd-service-upgrade’
but without traversing the service dependency graph to determine the
compilete set of obsolete services, no?  I feel that we should be
reusing ‘shepherd-service-upgrade’ or similar bits.  (I realize this is
already in ‘master’ for ‘guix deploy’, but since this is going to be
shared with ‘guix system’, we’d rather be extra cautious.)

Also, I think we should remove ‘false-if-exception’ around
‘unload-service’.

> +(define (install-bootloader-program installer-script bootcfg bootcfg-file target)
> +  "Return as a monadic value a derivation to build a scheme file that, upon
> +being evaluated, will install BOOTCFG to BOOTCFG-FILE, a target path, on
> +TARGET, a mount point, and subsequently run INSTALLER-SCRIPT."
> +  (gexp->script
> +   "install-bootloader.scm"
> +   (with-extensions (list guile-gcrypt)
> +     (with-imported-modules (source-module-closure '((gnu build install)
> +                                                     (guix store)
> +                                                     (guix utils)))
> +       #~(begin
> +           (use-modules (gnu build install)
> +                        (guix store)
> +                        (guix utils))
> +           (let* ((gc-root (string-append "/" %gc-roots-directory "/bootcfg"))
> +                  (temp-gc-root (string-append gc-root ".new")))
> +
> +             (switch-symlinks temp-gc-root gc-root)
> +
> +             (let ((installer-result
> +                    (false-if-exception
> +                     (begin
> +                       (install-boot-config #$bootcfg #$bootcfg-file #$target)
> +                       (with-output-to-string
> +                         (lambda ()
> +                           (primitive-load #$installer-script)))))))
> +               (unless installer-result
> +                 (delete-file temp-gc-root)
> +                 (error "failed to install bootloader"))
> +               (rename-file temp-gc-root gc-root)
> +               installer-result)))))))

I’d rather not swallow stdout and not use ‘error’.  Or at least, code
that runs ‘install-bootloader-program’ should be able to produce a
meaningful (and i18n’d) error message.  So the caller could do something
like:

  (define result
    (machine-eval #~(…
                     (guard (c ((message-condition? c)
                                (cons 'error (condition-message c))))
                       (invoke/quiet #$(install-bootloader-program …))
                       '(success)))
                  machine))

  (match result
    (('error message)
     (leave (G_ "failed to install bootloader:~%~a~%") message))
    (('success)
     #t))

Does that make sense?

That’s quite some boilerplate to the challenge will be to factorize it.

Ultimately, the code in (guix scripts system reconfigure) should be
parameterized by an evaluation procedure that would be either
‘machine-eval’ or some hypothetical ‘local-eval’ procedure to evaluate
things locally.

Thanks,
Ludo’.

  parent reply	other threads:[~2019-07-13 10:24 UTC|newest]

Thread overview: 52+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-07-08 19:52 [bug#36555] [PATCH 0/2] Refactor out common behavior for system reconfiguration Jakob L. Kreuze
2019-07-08 19:59 ` [bug#36555] [PATCH 1/2] guix system: Add 'reconfigure' module Jakob L. Kreuze
2019-07-08 20:01   ` [bug#36555] [PATCH 2/2] guix system: Reimplement 'reconfigure' Jakob L. Kreuze
2019-07-13 10:23   ` Ludovic Courtès [this message]
2019-07-13 17:44     ` [bug#36555] [PATCH 1/2] guix system: Add 'reconfigure' module Jakob L. Kreuze
2019-07-14 13:23       ` Ludovic Courtès
2019-07-15 15:36         ` Jakob L. Kreuze
2019-07-15 16:32           ` Ludovic Courtès
2019-07-15 23:57             ` Jakob L. Kreuze
2019-07-16 23:46               ` [bug#36555] [PATCH v3 0/3] Refactor out common behavior for system reconfiguration Jakob L. Kreuze
2019-07-16 23:47                 ` [bug#36555] [PATCH v3 1/3] guix system: Add 'reconfigure' module Jakob L. Kreuze
2019-07-16 23:48                   ` [bug#36555] [PATCH v3 2/3] guix system: Reimplement 'reconfigure' Jakob L. Kreuze
2019-07-16 23:48                     ` [bug#36555] [PATCH v3 3/3] tests: Add reconfigure system test Jakob L. Kreuze
2019-07-19 11:57                   ` [bug#36555] [PATCH v3 1/3] guix system: Add 'reconfigure' module Ludovic Courtès
2019-07-18 22:50                 ` [bug#36555] [PATCH v3 0/3] Refactor out common behavior for system reconfiguration Jakob L. Kreuze
2019-07-19 17:54                   ` [bug#36555] [PATCH v4 " Jakob L. Kreuze
2019-07-19 17:55                     ` [bug#36555] [PATCH v4 1/3] guix system: Add 'reconfigure' module Jakob L. Kreuze
2019-07-19 17:58                       ` [bug#36555] [PATCH v4 2/3] guix system: Reimplement 'reconfigure' Jakob L. Kreuze
2019-07-19 17:59                         ` [bug#36555] [PATCH v4 3/3] tests: Add reconfigure system test Jakob L. Kreuze
2019-07-20 14:50                           ` Ludovic Courtès
2019-07-22 18:16                             ` Jakob L. Kreuze
2019-07-22 18:23                               ` Jakob L. Kreuze
2019-07-22 18:54                               ` [bug#36555] [PATCH v5 0/3] Refactor out common behavior for system reconfiguration Jakob L. Kreuze
2019-07-22 18:56                                 ` [bug#36555] [PATCH v5 1/3] guix system: Add 'reconfigure' module Jakob L. Kreuze
2019-07-22 18:57                                   ` [bug#36555] [PATCH v5 2/3] guix system: Reimplement 'reconfigure' Jakob L. Kreuze
2019-07-22 18:57                                     ` [bug#36555] [PATCH v5 3/3] tests: Add reconfigure system test Jakob L. Kreuze
2019-07-23 22:30                                     ` [bug#36555] [PATCH v5 2/3] guix system: Reimplement 'reconfigure' Ludovic Courtès
2019-07-24  0:06                                       ` Jakob L. Kreuze
2019-07-24  0:48                                         ` Jakob L. Kreuze
2019-07-24 16:33                                           ` [bug#36555] [PATCH v6 0/3] Refactor out common behavior for system reconfiguration Jakob L. Kreuze
2019-07-24 16:34                                             ` [bug#36555] [PATCH v6 1/3] guix system: Add 'reconfigure' module Jakob L. Kreuze
2019-07-24 16:34                                               ` [bug#36555] [PATCH v6 2/3] guix system: Reimplement 'reconfigure' Jakob L. Kreuze
2019-07-24 16:35                                                 ` [bug#36555] [PATCH v6 3/3] tests: Add reconfigure system test Jakob L. Kreuze
2019-07-26 16:59                                                   ` bug#36555: " Ludovic Courtès
2019-07-26 17:53                                                     ` [bug#36555] " Jakob L. Kreuze
2019-07-24 22:46                                           ` [bug#36555] [PATCH v5 2/3] guix system: Reimplement 'reconfigure' Ludovic Courtès
2019-07-23 21:47                               ` [bug#36555] [PATCH v4 3/3] tests: Add reconfigure system test Ludovic Courtès
2019-07-24  0:01                                 ` Jakob L. Kreuze
2019-07-24 22:44                                   ` Ludovic Courtès
2019-07-20 14:40                         ` [bug#36555] [PATCH v4 2/3] guix system: Reimplement 'reconfigure' Ludovic Courtès
2019-07-20 14:29                       ` [bug#36555] [PATCH v4 1/3] guix system: Add 'reconfigure' module Ludovic Courtès
2019-07-30 16:55                         ` Jakob L. Kreuze
2019-08-23 21:00                           ` Ludovic Courtès
2019-07-19 17:56                     ` Jakob L. Kreuze
2019-07-19 19:36                   ` [bug#36555] [PATCH v3 0/3] Refactor out common behavior for system reconfiguration Christopher Lemmer Webber
2019-07-22 16:18                     ` Jakob L. Kreuze
2019-07-22 16:39                       ` Christopher Lemmer Webber
2019-07-09 13:26 ` [bug#36555] [PATCH 0/2] " Christopher Lemmer Webber
2019-07-09 19:07   ` [bug#36555] [PATCH v2 0/3] " Jakob L. Kreuze
2019-07-09 19:08     ` [bug#36555] [PATCH v2 1/3] guix system: Add 'reconfigure' module Jakob L. Kreuze
2019-07-09 19:09       ` [bug#36555] [PATCH v2 2/3] guix system: Reimplement 'reconfigure' Jakob L. Kreuze
2019-07-09 19:09         ` [bug#36555] [PATCH v2 3/3] tests: Add reconfigure system test Jakob L. Kreuze

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

  List information: https://guix.gnu.org/

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=87y3129qsn.fsf@gnu.org \
    --to=ludo@gnu.org \
    --cc=36555@debbugs.gnu.org \
    --cc=zerodaysfordays@sdf.lonestar.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
Code repositories for project(s) associated with this public inbox

	https://git.savannah.gnu.org/cgit/guix.git

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).