[PATCH 0/3] Defaulting to ci.guix.info (aka. berlin.guixsd.org)

  • Done
  • quality assurance status badge
Details
11 participants
  • Andreas Enge
  • Chris Marusich
  • Giovanni Biscuolo
  • Hartmut Goebel
  • Hartmut Goebel
  • Ludovic Courtès
  • Pierre Neidhardt
  • Marius Bakke
  • Meiyo Peng
  • Mark H Weaver
  • Ricardo Wurmus
Owner
unassigned
Submitted by
Ludovic Courtès
Severity
normal
L
L
Ludovic Courtès wrote on 3 Dec 2018 16:43
(address . guix-patches@gnu.org)
20181203154335.10366-1-ludo@gnu.org
Hello Guix!

These patches (actually the last one) switch Guix to default to
https://ci.guix.info for substitutes, in preparation for the
upcoming 0.16.0 release (hopefully this week!).

Rationale:

• berlin.guixsd.org generally performs better than hydra.gnu.org;

• berlin supports x86, ARMv7, and AArch64 (hydra lacks AArch64);

• berlin has much more storage space than hydra;

• berlin runs Cuirass on GuixSD, so software-wise, it corresponds
to what we want to push; Cuirass still has limitations in some
ways compared to Hydra, but we’re getting there, and
hydra.gnu.org is still up and running so we can still use it,
e.g., to track build progress until Cuirass is as convenient.

For the domain name I initially wanted “ci.guix.gnu.org” but we failed
to set that up. Oh well, I think that’s OK.

This change modifies the manual. Translations will thus be stale;
Julien, do you think it’d be reasonable to push a new pre-release
to the Translation Project? Or should be just live with the
slight inaccuracy?

Thanks,
Ludo’.

Ludovic Courtès (3):
etc: Add "ci.guix.info.pub" public key file.
Remove most references to hydra.gnu.org.
build: Default to https://ci.guix.infofor substitutes.

Makefile.am | 5 +-
build-aux/check-available-binaries.scm | 4 +-
.../check-final-inputs-self-contained.scm | 2 +-
config-daemon.ac | 10 +--
doc/contributing.texi | 2 +-
doc/guix.texi | 67 +++++++++----------
etc/substitutes/ci.guix.info.pub | 1 +
gnu/services/base.scm | 4 +-
gnu/system/install.scm | 2 +-
guix/scripts/build.scm | 2 +-
guix/scripts/size.scm | 2 +-
guix/scripts/substitute.scm | 2 +-
guix/self.scm | 3 +
guix/store.scm | 2 +-
14 files changed, 51 insertions(+), 57 deletions(-)
create mode 120000 etc/substitutes/ci.guix.info.pub

--
2.19.2
L
L
Ludovic Courtès wrote on 3 Dec 2018 16:53
[PATCH 3/3] build: Default to https://ci.guix.info for substitutes.
(address . 33600@debbugs.gnu.org)
20181203155335.10629-3-ludo@gnu.org
* config-daemon.ac (guix_substitute_urls): Always default to
"https://ci.guix.info".
* doc/guix.texi (SUBSTITUTE-SERVER): Switch to ci.guix.info.
* guix/scripts/build.scm (%default-log-urls): Likewise.
* guix/scripts/substitute.scm (%default-substitute-urls): Likewise.
* guix/store.scm (%default-substitute-urls): Likewise.
---
config-daemon.ac | 10 +---------
doc/guix.texi | 13 ++++---------
guix/scripts/build.scm | 2 +-
guix/scripts/substitute.scm | 2 +-
guix/store.scm | 2 +-
5 files changed, 8 insertions(+), 21 deletions(-)

Toggle diff (93 lines)
diff --git a/config-daemon.ac b/config-daemon.ac
index aac46817f5..e5b0ee082c 100644
--- a/config-daemon.ac
+++ b/config-daemon.ac
@@ -120,15 +120,7 @@ if test "x$guix_build_daemon" = "xyes"; then
dnl Determine the appropriate default list of substitute URLs (GnuTLS
dnl is required so we can default to 'https'.)
- case "x$host_cpu" in
- xaarch64)
- # Currently only berlin.guixsd.org provides aarch64 binaries.
- guix_substitute_urls="https://berlin.guixsd.org"
- ;;
- *)
- guix_substitute_urls="https://mirror.hydra.gnu.org"
- ;;
- esac
+ guix_substitute_urls="https://ci.guix.info"
AC_MSG_CHECKING([for default substitute URLs])
AC_MSG_RESULT([$guix_substitute_urls])
diff --git a/doc/guix.texi b/doc/guix.texi
index d96c35eace..a5ae1c65e6 100644
--- a/doc/guix.texi
+++ b/doc/guix.texi
@@ -14,7 +14,7 @@
@set KEY-SERVER pool.sks-keyservers.net
@c The official substitute server used by default.
-@set SUBSTITUTE-SERVER ci.guix.info
+@set SUBSTITUTE-SERVER mirror.hydra.gnu.org
@copying
Copyright @copyright{} 2012, 2013, 2014, 2015, 2016, 2017, 2018 Ludovic Courtès@*
@@ -2374,14 +2374,9 @@ Then, you can run something like this:
@end example
@quotation Note
-Similarly, the @file{berlin.guixsd.org.pub} file contains the public key
-for the project's new build farm, reachable at
-@indicateurl{https://berlin.guixsd.org}.
-
-As of this writing @code{berlin.guixsd.org} is being upgraded so it can
-better scale up, but you might want to give it a try. It is backed by
-20 x86_64/i686 build nodes and may be able to provide substitutes more
-quickly than @code{mirror.hydra.gnu.org}.
+Similarly, the @file{hydra.gnu.org.pub} file contains the public key
+of an independent build farm also run by the project, reachable at
+@indicateurl{https://mirror.hydra.gnu.org}.
@end quotation
Once this is in place, the output of a command like @code{guix build}
diff --git a/guix/scripts/build.scm b/guix/scripts/build.scm
index 5532c65eb6..0b7da3189e 100644
--- a/guix/scripts/build.scm
+++ b/guix/scripts/build.scm
@@ -65,7 +65,7 @@
(define %default-log-urls
;; Default base URLs for build logs.
- '("http://hydra.gnu.org/log"))
+ '("http://ci.guix.info/log"))
;; XXX: The following procedure cannot be in (guix store) because of the
;; dependency on (guix derivations).
diff --git a/guix/scripts/substitute.scm b/guix/scripts/substitute.scm
index eb82224016..d6dc9b6448 100755
--- a/guix/scripts/substitute.scm
+++ b/guix/scripts/substitute.scm
@@ -1052,7 +1052,7 @@ found."
(#f
;; This can only happen when this script is not invoked by the
;; daemon.
- '("http://hydra.gnu.org"))))
+ '("http://ci.guix.info"))))
(define substitute-urls
;; List of substitute URLs.
diff --git a/guix/store.scm b/guix/store.scm
index 9dc651b26c..509fd4def6 100644
--- a/guix/store.scm
+++ b/guix/store.scm
@@ -693,7 +693,7 @@ encoding conversion errors."
(map (if (false-if-exception (resolve-interface '(gnutls)))
(cut string-append "https://" <>)
(cut string-append "http://" <>))
- '("mirror.hydra.gnu.org")))
+ '("ci.guix.info")))
(define* (set-build-options server
#:key keep-failed? keep-going? fallback?
--
2.19.2
L
L
Ludovic Courtès wrote on 3 Dec 2018 16:53
[PATCH 1/3] etc: Add "ci.guix.info.pub" public key file.
(address . 33600@debbugs.gnu.org)
20181203155335.10629-1-ludo@gnu.org
* etc/substitutes/ci.guix.info.pub: New file.
* Makefile.am (dist_pkgdata_DATA): Add it.
* guix/self.scm (miscellaneous-files): Add
"share/guix/ci.guix.info.pub".
---
Makefile.am | 3 ++-
etc/substitutes/ci.guix.info.pub | 1 +
guix/self.scm | 3 +++
3 files changed, 6 insertions(+), 1 deletion(-)
create mode 120000 etc/substitutes/ci.guix.info.pub

Toggle diff (38 lines)
diff --git a/Makefile.am b/Makefile.am
index 32cebd5917..6803666a57 100644
--- a/Makefile.am
+++ b/Makefile.am
@@ -459,7 +459,8 @@ check-system: $(GOBJECTS)
# Public key used to sign substitutes from hydra.gnu.org & co.
dist_pkgdata_DATA = \
etc/substitutes/hydra.gnu.org.pub \
- etc/substitutes/berlin.guixsd.org.pub
+ etc/substitutes/berlin.guixsd.org.pub \
+ etc/substitutes/ci.guix.info.pub
# Bash completion file.
dist_bashcompletion_DATA = etc/completion/bash/guix \
diff --git a/etc/substitutes/ci.guix.info.pub b/etc/substitutes/ci.guix.info.pub
new file mode 120000
index 0000000000..b1d1e4d763
--- /dev/null
+++ b/etc/substitutes/ci.guix.info.pub
@@ -0,0 +1 @@
+berlin.guixsd.org.pub
\ No newline at end of file
diff --git a/guix/self.scm b/guix/self.scm
index 8476c422ec..f2db3dbf52 100644
--- a/guix/self.scm
+++ b/guix/self.scm
@@ -428,6 +428,9 @@ load path."
,(file-append* source
"/etc/substitutes/hydra.gnu.org.pub"))
("share/guix/berlin.guixsd.org.pub"
+ ,(file-append* source
+ "/etc/substitutes/berlin.guixsd.org.pub"))
+ ("share/guix/ci.guix.info.pub" ;alias
,(file-append* source "/etc/substitutes/berlin.guixsd.org.pub")))))
(define* (whole-package name modules dependencies
--
2.19.2
L
L
Ludovic Courtès wrote on 3 Dec 2018 16:53
[PATCH 2/3] Remove most references to hydra.gnu.org.
(address . 33600@debbugs.gnu.org)
20181203155335.10629-2-ludo@gnu.org
* Makefile.am (assert-binaries-available): Don't mention hydra.gnu.org
in comment.
* build-aux/check-available-binaries.scm: Likewise.
* build-aux/check-final-inputs-self-contained.scm: Likewise.
* doc/guix.texi (SUBSTITUTE-SERVER): New variable.
Use it throughout instead of "mirror.hydra.gnu.org".
* doc/contributing.texi (Submitting Patches): Likewise.
* gnu/services/base.scm (hydra-key-authorization)
(guix-activation): Remove mentions of "hydra.gnu.org" in comments and
messages.
* gnu/system/install.scm (%installation-services): Likewise.
* guix/scripts/size.scm (guix-size): Likewise.
---
Makefile.am | 2 +-
build-aux/check-available-binaries.scm | 4 +-
.../check-final-inputs-self-contained.scm | 2 +-
doc/contributing.texi | 2 +-
doc/guix.texi | 56 ++++++++++---------
gnu/services/base.scm | 4 +-
gnu/system/install.scm | 2 +-
guix/scripts/size.scm | 2 +-
8 files changed, 38 insertions(+), 36 deletions(-)

Toggle diff (310 lines)
diff --git a/Makefile.am b/Makefile.am
index 6803666a57..a7a67e81cf 100644
--- a/Makefile.am
+++ b/Makefile.am
@@ -763,7 +763,7 @@ assert-no-store-file-names: $(distdir)/ChangeLog
exit 1 ; \
fi
-# Make sure hydra.gnu.org has the important binaries.
+# Make sure important substitutes are available.
assert-binaries-available: $(GOBJECTS)
$(AM_V_at)$(top_builddir)/pre-inst-env "$(GUILE)" \
"$(top_srcdir)/build-aux/check-available-binaries.scm"
diff --git a/build-aux/check-available-binaries.scm b/build-aux/check-available-binaries.scm
index b832d99935..c80db1a37e 100644
--- a/build-aux/check-available-binaries.scm
+++ b/build-aux/check-available-binaries.scm
@@ -1,5 +1,5 @@
;;; GNU Guix --- Functional package management for GNU
-;;; Copyright © 2013, 2014, 2015, 2016, 2017 Ludovic Courtès <ludo@gnu.org>
+;;; Copyright © 2013, 2014, 2015, 2016, 2017, 2018 Ludovic Courtès <ludo@gnu.org>
;;;
;;; This file is part of GNU Guix.
;;;
@@ -17,7 +17,7 @@
;;; along with GNU Guix. If not, see <http://www.gnu.org/licenses/>.
;;;
-;;; Check whether important binaries are available at hydra.gnu.org.
+;;; Check whether important binaries are available.
;;;
(use-modules (guix store)
diff --git a/build-aux/check-final-inputs-self-contained.scm b/build-aux/check-final-inputs-self-contained.scm
index 37dc883d3c..c819086508 100644
--- a/build-aux/check-final-inputs-self-contained.scm
+++ b/build-aux/check-final-inputs-self-contained.scm
@@ -17,7 +17,7 @@
;;; along with GNU Guix. If not, see <http://www.gnu.org/licenses/>.
;;;
-;;; Check whether important binaries are available at hydra.gnu.org.
+;;; Check whether important binaries are available.
;;;
(use-modules (guix store)
diff --git a/doc/contributing.texi b/doc/contributing.texi
index 7417f28d14..c55eb63382 100644
--- a/doc/contributing.texi
+++ b/doc/contributing.texi
@@ -443,7 +443,7 @@ as timestamps or randomly-generated output in the build result.
Another option is to use @command{guix challenge} (@pxref{Invoking guix
challenge}). You may run it once the package has been committed and
-built by @code{hydra.gnu.org} to check whether it obtains the same
+built by @code{@value{SUBSTITUTE-SERVER}} to check whether it obtains the same
result as you did. Better yet: Find another machine that can build it
and run @command{guix publish}. Since the remote build machine is
likely different from yours, this can catch non-determinism issues
diff --git a/doc/guix.texi b/doc/guix.texi
index b0aa5aa872..d96c35eace 100644
--- a/doc/guix.texi
+++ b/doc/guix.texi
@@ -13,6 +13,9 @@
@set OPENPGP-SIGNING-KEY-ID 3CE464558A84FDC69DB40CFB090B11993D9AEBB5
@set KEY-SERVER pool.sks-keyservers.net
+@c The official substitute server used by default.
+@set SUBSTITUTE-SERVER ci.guix.info
+
@copying
Copyright @copyright{} 2012, 2013, 2014, 2015, 2016, 2017, 2018 Ludovic Courtès@*
Copyright @copyright{} 2013, 2014, 2016 Andreas Enge@*
@@ -1274,8 +1277,7 @@ remote procedure call (@pxref{The Store}).
@anchor{daemon-substitute-urls}
Consider @var{urls} the default whitespace-separated list of substitute
source URLs. When this option is omitted,
-@indicateurl{https://mirror.hydra.gnu.org https://hydra.gnu.org} is used
-(@code{mirror.hydra.gnu.org} is a mirror of @code{hydra.gnu.org}).
+@indicateurl{https://@value{SUBSTITUTE-SERVER}} is used.
This means that substitutes may be downloaded from @var{urls}, as long
as they are signed by a trusted signature (@pxref{Substitutes}).
@@ -2321,7 +2323,7 @@ also result from derivation builds, can be available as substitutes.
@cindex hydra
@cindex build farm
-The @code{mirror.hydra.gnu.org} server is a front-end to an official build farm
+The @code{@value{SUBSTITUTE-SERVER}} server is a front-end to an official build farm
that builds packages from Guix continuously for some
architectures, and makes them available as substitutes. This is the
default source of substitutes; it can be overridden by passing the
@@ -2353,22 +2355,22 @@ other substitute server.
@cindex substitutes, authorization thereof
@cindex access control list (ACL), for substitutes
@cindex ACL (access control list), for substitutes
-To allow Guix to download substitutes from @code{hydra.gnu.org} or a
+To allow Guix to download substitutes from @code{@value{SUBSTITUTE-SERVER}} or a
mirror thereof, you
must add its public key to the access control list (ACL) of archive
imports, using the @command{guix archive} command (@pxref{Invoking guix
-archive}). Doing so implies that you trust @code{hydra.gnu.org} to not
+archive}). Doing so implies that you trust @code{@value{SUBSTITUTE-SERVER}} to not
be compromised and to serve genuine substitutes.
-The public key for @code{hydra.gnu.org} is installed along with Guix, in
-@code{@var{prefix}/share/guix/hydra.gnu.org.pub}, where @var{prefix} is
+The public key for @code{@value{SUBSTITUTE-SERVER}} is installed along with Guix, in
+@code{@var{prefix}/share/guix/@value{SUBSTITUTE-SERVER}.pub}, where @var{prefix} is
the installation prefix of Guix. If you installed Guix from source,
make sure you checked the GPG signature of
@file{guix-@value{VERSION}.tar.gz}, which contains this public key file.
Then, you can run something like this:
@example
-# guix archive --authorize < @var{prefix}/share/guix/hydra.gnu.org.pub
+# guix archive --authorize < @var{prefix}/share/guix/@value{SUBSTITUTE-SERVER}.pub
@end example
@quotation Note
@@ -2409,7 +2411,7 @@ $ guix build emacs --dry-run
@end example
@noindent
-This indicates that substitutes from @code{hydra.gnu.org} are usable and
+This indicates that substitutes from @code{@value{SUBSTITUTE-SERVER}} are usable and
will be downloaded, when possible, for future builds.
@cindex substitutes, how to disable
@@ -2501,9 +2503,9 @@ by a server.
Today, each individual's control over their own computing is at the
mercy of institutions, corporations, and groups with enough power and
determination to subvert the computing infrastructure and exploit its
-weaknesses. While using @code{hydra.gnu.org} substitutes can be
+weaknesses. While using @code{@value{SUBSTITUTE-SERVER}} substitutes can be
convenient, we encourage users to also build on their own, or even run
-their own build farm, such that @code{hydra.gnu.org} is less of an
+their own build farm, such that @code{@value{SUBSTITUTE-SERVER}} is less of an
interesting target. One way to help is by publishing the software you
build using @command{guix publish} so that others have one more choice
of server to download substitutes from (@pxref{Invoking guix publish}).
@@ -3658,11 +3660,11 @@ Read a single-item archive as served by substitute servers
low-level operation needed in only very narrow use cases; see below.
For example, the following command extracts the substitute for Emacs
-served by @code{hydra.gnu.org} to @file{/tmp/emacs}:
+served by @code{@value{SUBSTITUTE-SERVER}} to @file{/tmp/emacs}:
@example
$ wget -O - \
- https://hydra.gnu.org/nar/@dots{}-emacs-24.5 \
+ https://@value{SUBSTITUTE-SERVER}/nar/@dots{}-emacs-24.5 \
| bunzip2 | guix archive -x /tmp/emacs
@end example
@@ -6683,7 +6685,7 @@ but you are actually on an @code{x86_64} machine:
@example
$ guix build --log-file gdb -s mips64el-linux
-https://hydra.gnu.org/log/@dots{}-gdb-7.10
+https://@value{SUBSTITUTE-SERVER}/log/@dots{}-gdb-7.10
@end example
You can freely access a huge library of build logs!
@@ -8342,7 +8344,7 @@ When @command{guix publish} runs, it spawns an HTTP server which allows
anyone with network access to obtain substitutes from it. This means
that any machine running Guix can also act as if it were a build farm,
since the HTTP interface is compatible with Hydra, the software behind
-the @code{hydra.gnu.org} build farm.
+the @code{@value{SUBSTITUTE-SERVER}} build farm.
For security, each substitute is signed, allowing recipients to check
their authenticity and integrity (@pxref{Substitutes}). Because
@@ -8579,20 +8581,20 @@ any given store item.
The command output looks like this:
@smallexample
-$ guix challenge --substitute-urls="https://hydra.gnu.org https://guix.example.org"
-updating list of substitutes from 'https://hydra.gnu.org'... 100.0%
+$ guix challenge --substitute-urls="https://@value{SUBSTITUTE-SERVER} https://guix.example.org"
+updating list of substitutes from 'https://@value{SUBSTITUTE-SERVER}'... 100.0%
updating list of substitutes from 'https://guix.example.org'... 100.0%
/gnu/store/@dots{}-openssl-1.0.2d contents differ:
local hash: 0725l22r5jnzazaacncwsvp9kgf42266ayyp814v7djxs7nk963q
- https://hydra.gnu.org/nar/@dots{}-openssl-1.0.2d: 0725l22r5jnzazaacncwsvp9kgf42266ayyp814v7djxs7nk963q
+ https://@value{SUBSTITUTE-SERVER}/nar/@dots{}-openssl-1.0.2d: 0725l22r5jnzazaacncwsvp9kgf42266ayyp814v7djxs7nk963q
https://guix.example.org/nar/@dots{}-openssl-1.0.2d: 1zy4fmaaqcnjrzzajkdn3f5gmjk754b43qkq47llbyak9z0qjyim
/gnu/store/@dots{}-git-2.5.0 contents differ:
local hash: 00p3bmryhjxrhpn2gxs2fy0a15lnip05l97205pgbk5ra395hyha
- https://hydra.gnu.org/nar/@dots{}-git-2.5.0: 069nb85bv4d4a6slrwjdy8v1cn4cwspm3kdbmyb81d6zckj3nq9f
+ https://@value{SUBSTITUTE-SERVER}/nar/@dots{}-git-2.5.0: 069nb85bv4d4a6slrwjdy8v1cn4cwspm3kdbmyb81d6zckj3nq9f
https://guix.example.org/nar/@dots{}-git-2.5.0: 0mdqa9w1p6cmli6976v4wi0sw9r4p5prkj7lzfd1877wk11c9c73
/gnu/store/@dots{}-pius-2.1.1 contents differ:
local hash: 0k4v3m9z1zp8xzzizb7d8kjj72f9172xv078sq4wl73vnq9ig3ax
- https://hydra.gnu.org/nar/@dots{}-pius-2.1.1: 0k4v3m9z1zp8xzzizb7d8kjj72f9172xv078sq4wl73vnq9ig3ax
+ https://@value{SUBSTITUTE-SERVER}/nar/@dots{}-pius-2.1.1: 0k4v3m9z1zp8xzzizb7d8kjj72f9172xv078sq4wl73vnq9ig3ax
https://guix.example.org/nar/@dots{}-pius-2.1.1: 1cy25x1a4fzq5rk0pmvc8xhwyffnqz95h2bpvqsz2mpvlbccy0gs
@dots{}
@@ -8612,7 +8614,7 @@ the servers obtained a result different from the local build.
@cindex non-determinism, in package builds
As an example, @code{guix.example.org} always gets a different answer.
-Conversely, @code{hydra.gnu.org} agrees with local builds, except in the
+Conversely, @code{@value{SUBSTITUTE-SERVER}} agrees with local builds, except in the
case of Git. This might indicate that the build process of Git is
non-deterministic, meaning that its output varies as a function of
various things that Guix does not fully control, in spite of building
@@ -8626,14 +8628,14 @@ To find out what is wrong with this Git binary, we can do something along
these lines (@pxref{Invoking guix archive}):
@example
-$ wget -q -O - https://hydra.gnu.org/nar/@dots{}-git-2.5.0 \
+$ wget -q -O - https://@value{SUBSTITUTE-SERVER}/nar/@dots{}-git-2.5.0 \
| guix archive -x /tmp/git
$ diff -ur --no-dereference /gnu/store/@dots{}-git.2.5.0 /tmp/git
@end example
This command shows the difference between the files resulting from the
local build, and the files resulting from the build on
-@code{hydra.gnu.org} (@pxref{Overview, Comparing and Merging Files,,
+@code{@value{SUBSTITUTE-SERVER}} (@pxref{Overview, Comparing and Merging Files,,
diffutils, Comparing and Merging Files}). The @command{diff} command
works great for text files. When binary files differ, a better option
is @uref{https://diffoscope.org/, Diffoscope}, a tool that helps
@@ -8648,7 +8650,7 @@ In the meantime, @command{guix challenge} is one tool to help address
the problem.
If you are writing packages for Guix, you are encouraged to check
-whether @code{hydra.gnu.org} and other substitute servers obtain the
+whether @code{@value{SUBSTITUTE-SERVER}} and other substitute servers obtain the
same build result as you did with:
@example
@@ -11238,14 +11240,14 @@ Number of build user accounts to create.
@item @code{authorize-key?} (default: @code{#t})
@cindex substitutes, authorization thereof
Whether to authorize the substitute keys listed in
-@code{authorized-keys}---by default that of @code{hydra.gnu.org}
+@code{authorized-keys}---by default that of @code{@value{SUBSTITUTE-SERVER}}
(@pxref{Substitutes}).
@vindex %default-authorized-guix-keys
@item @code{authorized-keys} (default: @var{%default-authorized-guix-keys})
The list of authorized key files for archive imports, as a list of
string-valued gexps (@pxref{Invoking guix archive}). By default, it
-contains that of @code{hydra.gnu.org} (@pxref{Substitutes}).
+contains that of @code{@value{SUBSTITUTE-SERVER}} (@pxref{Substitutes}).
@item @code{use-substitutes?} (default: @code{#t})
Whether to use substitutes.
@@ -23747,7 +23749,7 @@ system}.
@cindex substituter
Users can obtain the new package definition simply by running
@command{guix pull} (@pxref{Invoking guix pull}). When
-@code{hydra.gnu.org} is done building the package, installing the
+@code{@value{SUBSTITUTE-SERVER}} is done building the package, installing the
package automatically downloads binaries from there
(@pxref{Substitutes}). The only place where human intervention is
needed is to review and apply the patch.
diff --git a/gnu/services/base.scm b/gnu/services/base.scm
index 228d3c5926..cee9898d79 100644
--- a/gnu/services/base.scm
+++ b/gnu/services/base.scm
@@ -1518,7 +1518,7 @@ public key, with GUIX."
(let ((status (cdr (waitpid pid))))
(unless (zero? status)
(format (current-error-port) "warning: \
-failed to register hydra.gnu.org public key: ~a~%" status))))))))
+failed to register public key '~a': ~a~%" key status))))))))
(define %default-authorized-guix-keys
;; List of authorized substitute keys.
@@ -1630,7 +1630,7 @@ failed to register hydra.gnu.org public key: ~a~%" status))))))))
;; otherwise call 'chown' here, but the problem is that on a COW overlayfs,
;; chown leads to an entire copy of the tree, which is a bad idea.
- ;; Optionally authorize hydra.gnu.org's key.
+ ;; Optionally authorize substitute server keys.
(if authorize-key?
#~(begin
#$@(map (cut hydra-key-authorization <> guix) keys))
diff --git a/gnu/system/install.scm b/gnu/system/install.scm
index 45b3a0c839..c345ba0626 100644
--- a/gnu/system/install.scm
+++ b/gnu/system/install.scm
@@ -250,7 +250,7 @@ You have been warned. Thanks for being so brave.\x1b[0m
;; The usual services.
(syslog-service)
- ;; The build daemon. Register the hydra.gnu.org key as trusted.
+ ;; The build daemon. Register the official server keys as trusted.
;; This allows the installation process to use substitutes by
;; default.
(guix-service (guix-configuration (authorize-key? #t)))
diff --git a/guix/scripts/size.scm b/guix/scripts/size.scm
index 344be40883..25218a2945 100644
--- a/guix/scripts/size.scm
+++ b/guix/scripts/size.scm
@@ -297,7 +297,7 @@ Report the size of PACKAGE and its dependencies.\n"))
(leave (G_ "missing store item argument\n")))
((files ..1)
(leave-on-EPIPE
- ;; Turn off grafts because (1) hydra.gnu.org does not serve grafted
+ ;; Turn off grafts because (1) substitute servers do not serve grafted
;; packages, and (2) they do not make any difference on the
;; resulting size.
(parameterize ((%graft? #f))
--
2.19.2
L
L
Ludovic Courtès wrote on 3 Dec 2018 17:12
Using a CDN or some other mirror?
(address . 33600@debbugs.gnu.org)
87tvju6145.fsf@gnu.org
Hello,

Ludovic Courtès <ludo@gnu.org> skribis:

Toggle quote (4 lines)
> These patches (actually the last one) switch Guix to default to
> <https://ci.guix.info> for substitutes, in preparation for the
> upcoming 0.16.0 release (hopefully this week!).

Right now, ci.guix.info points to berlin.guixsd.org, the front-end of
the build farm hosted at the MDC.

The previous setup was that mirror.hydra.gnu.org mirrors hydra.gnu.org
(the actual build farm front-end) using an nginx proxy:


This provides a bit of redundancy that we don’t have currently for
berlin.

Thus, I’m thinking about using a similar setup, but hosting the mirror
on some Big Corp CDN or similar. Chris Marusich came up with a setup
along these lines a while back:


Compared to Chris’s setup, given that ‘guix publish’ now provides
‘Cache-Control’ headers (that was not the case back then, see
caching in the proxy should Just Work.

I would like us to set up such a mirror for berlin and then have
ci.guix.info point to that. The project should be able to pay the
hosting fees.

Thoughts?

Would someone like to get started? You’ll undoubtedly get all the
appreciation of each one of us and a beverage of your choice next time
we meet! :-)

Thanks,
Ludo’.
R
R
Ricardo Wurmus wrote on 3 Dec 2018 21:44
Re: [PATCH 1/3] etc: Add "ci.guix.info.pub" public key file.
(name . Ludovic Courtès)(address . ludo@gnu.org)(address . 33600@debbugs.gnu.org)
87y396gx1e.fsf@elephly.net
Ludovic Courtès <ludo@gnu.org> writes:

Toggle quote (5 lines)
> * etc/substitutes/ci.guix.info.pub: New file.
> * Makefile.am (dist_pkgdata_DATA): Add it.
> * guix/self.scm (miscellaneous-files): Add
> "share/guix/ci.guix.info.pub".

LGTM!

--
Ricardo
R
R
Ricardo Wurmus wrote on 3 Dec 2018 21:47
Re: Using a CDN or some other mirror?
(name . Ludovic Courtès)(address . ludo@gnu.org)
87wooqgwx3.fsf@elephly.net
Ludovic Courtès <ludo@gnu.org> writes:

Toggle quote (25 lines)
> Hello,
>
> Ludovic Courtès <ludo@gnu.org> skribis:
>
>> These patches (actually the last one) switch Guix to default to
>> <https://ci.guix.info> for substitutes, in preparation for the
>> upcoming 0.16.0 release (hopefully this week!).
>
> Right now, ci.guix.info points to berlin.guixsd.org, the front-end of
> the build farm hosted at the MDC.
>
> The previous setup was that mirror.hydra.gnu.org mirrors hydra.gnu.org
> (the actual build farm front-end) using an nginx proxy:
>
> https://git.savannah.gnu.org/cgit/guix/maintenance.git/tree/hydra/nginx/mirror.conf
>
> This provides a bit of redundancy that we don’t have currently for
> berlin.
>
> Thus, I’m thinking about using a similar setup, but hosting the mirror
> on some Big Corp CDN or similar. Chris Marusich came up with a setup
> along these lines a while back:
>
> https://lists.gnu.org/archive/html/guix-devel/2016-03/msg00312.html

Large ISPs also provide CDN services. I already contacted Deutsche
Telekom so that we can compare their CDN offer with the Amazon Cloudfont
setup that Chris has configured.

--
Ricardo
M
M
Mark H Weaver wrote on 4 Dec 2018 00:44
Re: [PATCH 0/3] Defaulting to ci.guix.info (aka. berlin.guixsd.org)
(name . Ludovic Courtès)(address . ludo@gnu.org)
8736re41kp.fsf@netris.org
Hi Ludovic,

Ludovic Courtès <ludo@gnu.org> writes:

Toggle quote (4 lines)
> These patches (actually the last one) switch Guix to default to
> <https://ci.guix.info> for substitutes, in preparation for the
> upcoming 0.16.0 release (hopefully this week!).

Who owns the guix.info domain? Also, who owns the guixsd.org domain?

Thanks,
Mark
R
R
Ricardo Wurmus wrote on 4 Dec 2018 06:55
(name . Mark H Weaver)(address . mhw@netris.org)
87va49hm4d.fsf@elephly.net
Hi Mark,

Toggle quote (8 lines)
> Ludovic Courtès <ludo@gnu.org> writes:
>
>> These patches (actually the last one) switch Guix to default to
>> <https://ci.guix.info> for substitutes, in preparation for the
>> upcoming 0.16.0 release (hopefully this week!).
>
> Who owns the guix.info domain?

I registered it and offered it to Guix Europe (though we aren’t yet
sharing administration of the domain).

--
Ricardo
R
R
Ricardo Wurmus wrote on 3 Dec 2018 21:44
Re: [PATCH 2/3] Remove most references to hydra.gnu.org.
(name . Ludovic Courtès)(address . ludo@gnu.org)
87zhtmgx2n.fsf@elephly.net
Ludovic Courtès <ludo@gnu.org> writes:

Toggle quote (13 lines)
> * Makefile.am (assert-binaries-available): Don't mention hydra.gnu.org
> in comment.
> * build-aux/check-available-binaries.scm: Likewise.
> * build-aux/check-final-inputs-self-contained.scm: Likewise.
> * doc/guix.texi (SUBSTITUTE-SERVER): New variable.
> Use it throughout instead of "mirror.hydra.gnu.org".
> * doc/contributing.texi (Submitting Patches): Likewise.
> * gnu/services/base.scm (hydra-key-authorization)
> (guix-activation): Remove mentions of "hydra.gnu.org" in comments and
> messages.
> * gnu/system/install.scm (%installation-services): Likewise.
> * guix/scripts/size.scm (guix-size): Likewise.

This looks good to me.

--
Ricardo
A
A
Andreas Enge wrote on 4 Dec 2018 09:59
Re: [bug#33600] [PATCH 0/3] Defaulting to ci.guix.info (aka. berlin.guixsd.org)
(name . Mark H Weaver)(address . mhw@netris.org)
20181204085948.GA1690@jurong
On Mon, Dec 03, 2018 at 06:44:43PM -0500, Mark H Weaver wrote:
Toggle quote (2 lines)
> Who owns the guix.info domain? Also, who owns the guixsd.org domain?

guixsd.org is owned by the Guix Europe association. guix.info, as far as
I know, by Ricardo. Plans are to drop guixsd.org in the long run, and I
would suggest to transfer guix.info to Guix Europe at the next renewal.

Andreas
L
L
Ludovic Courtès wrote on 4 Dec 2018 10:03
(name . Ricardo Wurmus)(address . rekado@elephly.net)
87bm617jga.fsf@gnu.org
Hello,

Ricardo Wurmus <rekado@elephly.net> skribis:

Toggle quote (11 lines)
>> Ludovic Courtès <ludo@gnu.org> writes:
>>
>>> These patches (actually the last one) switch Guix to default to
>>> <https://ci.guix.info> for substitutes, in preparation for the
>>> upcoming 0.16.0 release (hopefully this week!).
>>
>> Who owns the guix.info domain?
>
> I registered it and offered it to Guix Europe (though we aren’t yet
> sharing administration of the domain).

Could you make sure several people have access to the administration
interface?

As for guixsd.org, it was initially owned by Andreas Enge who
transferred it to Guix Europe and also provided credentials to me and I
think a couple of other people, though I forgot who. Andreas?

We should write all this information down in guix-maintenance.git like
we did for build machines, so people know who to contact in case
something goes wrong or what to do when someone leaves the project.

Thanks,
Ludo’.
A
A
Andreas Enge wrote on 4 Dec 2018 11:08
(name . Ludovic Courtès)(address . ludo@gnu.org)
20181204100807.GA4134@jurong
On Tue, Dec 04, 2018 at 10:03:01AM +0100, Ludovic Courtï¿œs wrote:
Toggle quote (4 lines)
> As for guixsd.org, it was initially owned by Andreas Enge who
> transferred it to Guix Europe and also provided credentials to me and I
> think a couple of other people, though I forgot who. Andreas?

Apart from you, to Ricardo Wurmus, Christopher Baines and Manolis Ragkousis.
I am of course happy to share the credentials with more people from the
Guix sysadmin group.

Andreas
L
L
Ludovic Courtès wrote on 4 Dec 2018 11:21
Re: [bug#33600] [PATCH 2/3] Remove most references to hydra.gnu.org.
(name . Ricardo Wurmus)(address . rekado@elephly.net)
87mupl6188.fsf@gnu.org
Ricardo Wurmus <rekado@elephly.net> skribis:

Toggle quote (17 lines)
> Ludovic Courtès <ludo@gnu.org> writes:
>
>> * Makefile.am (assert-binaries-available): Don't mention hydra.gnu.org
>> in comment.
>> * build-aux/check-available-binaries.scm: Likewise.
>> * build-aux/check-final-inputs-self-contained.scm: Likewise.
>> * doc/guix.texi (SUBSTITUTE-SERVER): New variable.
>> Use it throughout instead of "mirror.hydra.gnu.org".
>> * doc/contributing.texi (Submitting Patches): Likewise.
>> * gnu/services/base.scm (hydra-key-authorization)
>> (guix-activation): Remove mentions of "hydra.gnu.org" in comments and
>> messages.
>> * gnu/system/install.scm (%installation-services): Likewise.
>> * guix/scripts/size.scm (guix-size): Likewise.
>
> This looks good to me.

Great, pushed!

Ludo’.
Closed
L
L
Ludovic Courtès wrote on 4 Dec 2018 11:28
Re: [bug#33600] [PATCH 0/3] Defaulting to ci.guix.info (aka. berlin.guixsd.org)
(name . Andreas Enge)(address . andreas@enge.fr)
87in0960ws.fsf@gnu.org
Hello,

Andreas Enge <andreas@enge.fr> skribis:

Toggle quote (7 lines)
> On Mon, Dec 03, 2018 at 06:44:43PM -0500, Mark H Weaver wrote:
>> Who owns the guix.info domain? Also, who owns the guixsd.org domain?
>
> guixsd.org is owned by the Guix Europe association. guix.info, as far as
> I know, by Ricardo. Plans are to drop guixsd.org in the long run, and I
> would suggest to transfer guix.info to Guix Europe at the next renewal.

Awesome. I’ve added a file to record that info in
guix-maintenance.git. Please feel free to complete it; for instance
it’d be good to know the expiration date of each domain.

Thanks!

Ludo’.
A
A
Andreas Enge wrote on 4 Dec 2018 11:46
(name . Ludovic Courtès)(address . ludo@gnu.org)
20181204104659.GA6125@jurong
On Tue, Dec 04, 2018 at 11:28:51AM +0100, Ludovic Courtès wrote:
Toggle quote (4 lines)
> Awesome. I’ve added a file to record that info in
> guix-maintenance.git. Please feel free to complete it; for instance
> it’d be good to know the expiration date of each domain.

Done for guixsd.org. It expires on January 27. I would be okay to let it
run out if guix.info is sufficiently marketed in the meantime.

Andreas
L
L
Ludovic Courtès wrote on 4 Dec 2018 15:12
(name . Andreas Enge)(address . andreas@enge.fr)
87in091iuz.fsf@gnu.org
Andreas Enge <andreas@enge.fr> skribis:

Toggle quote (8 lines)
> On Tue, Dec 04, 2018 at 11:28:51AM +0100, Ludovic Courtès wrote:
>> Awesome. I’ve added a file to record that info in
>> guix-maintenance.git. Please feel free to complete it; for instance
>> it’d be good to know the expiration date of each domain.
>
> Done for guixsd.org. It expires on January 27. I would be okay to let it
> run out if guix.info is sufficiently marketed in the meantime.

Thanks. I think it’s safer to keep it for a while because there may be
machine referring to berlin.guixsd.org here and there.

Ludo’.
L
L
Ludovic Courtès wrote on 4 Dec 2018 15:14
Re: [PATCH 3/3] build: Default to https://ci.guix.info for substitutes.
(name . Ricardo Wurmus)(address . rekado@elephly.net)
87a7ll1irc.fsf@gnu.org
Ricardo Wurmus <rekado@elephly.net> skribis:

Toggle quote (17 lines)
> Ludovic Courtès <ludo@gnu.org> writes:
>
>> * doc/guix.texi (SUBSTITUTE-SERVER): Switch to ci.guix.info.
> […]
>> diff --git a/doc/guix.texi b/doc/guix.texi
>> index d96c35eace..a5ae1c65e6 100644
>> --- a/doc/guix.texi
>> +++ b/doc/guix.texi
>> @@ -14,7 +14,7 @@
>> @set KEY-SERVER pool.sks-keyservers.net
>>
>> @c The official substitute server used by default.
>> -@set SUBSTITUTE-SERVER ci.guix.info
>> +@set SUBSTITUTE-SERVER mirror.hydra.gnu.org
>
> Is this really correct?

Oops, I must have messed up while preparing the patches.
Thanks for the heads-up!

Ludo’.
R
R
Ricardo Wurmus wrote on 3 Dec 2018 21:38
(name . Ludovic Courtès)(address . ludo@gnu.org)
871s6yibw3.fsf@elephly.net
Ludovic Courtès <ludo@gnu.org> writes:

Toggle quote (1 lines)
> * doc/guix.texi (SUBSTITUTE-SERVER): Switch to ci.guix.info.
[…]
Toggle quote (11 lines)
> diff --git a/doc/guix.texi b/doc/guix.texi
> index d96c35eace..a5ae1c65e6 100644
> --- a/doc/guix.texi
> +++ b/doc/guix.texi
> @@ -14,7 +14,7 @@
> @set KEY-SERVER pool.sks-keyservers.net
>
> @c The official substitute server used by default.
> -@set SUBSTITUTE-SERVER ci.guix.info
> +@set SUBSTITUTE-SERVER mirror.hydra.gnu.org

Is this really correct?

--
Ricardo
C
C
Chris Marusich wrote on 9 Dec 2018 04:33
Re: Using a CDN or some other mirror?
87ftv7l6gy.fsf@gmail.com
Hi everyone,

ludo@gnu.org (Ludovic Courtès) writes:

Toggle quote (19 lines)
> Ludovic Courtès <ludo@gnu.org> skribis:
>
> [...] I’m thinking about using a similar setup, but hosting the mirror
> on some Big Corp CDN or similar. Chris Marusich came up with a setup
> along these lines a while back:
>
> https://lists.gnu.org/archive/html/guix-devel/2016-03/msg00312.html
>
> Compared to Chris’s setup, given that ‘guix publish’ now provides
> ‘Cache-Control’ headers (that was not the case back then, see
> <https://lists.gnu.org/archive/html/guix-devel/2016-03/msg00360.html>),
> caching in the proxy should Just Work.
>
> I would like us to set up such a mirror for berlin and then have
> ci.guix.info point to that. The project should be able to pay the
> hosting fees.
>
> Thoughts?

Regarding DNS, it would be nice if we could use an official GNU
subdomain. If we can't use a GNU subdomain, we should at least make
sure we have some kind of DNS auto-renewal set up so that nobody can
poach our domain names. And the operators should take appropriate
precautions when sharing any credentials used for managing it all.

Regarding CDNs, I definitely think it's worth a try! Even Debian is
using CloudFront (cloudfront.debian.net). In fact, email correspondence
suggests that as of 2013, Amazon may even have been paying for it!


I wonder if Amazon would be willing to pay for our CloudFront
distribution if we asked them nicely?

In any case, before deciding to use Amazon CloudFront for ci.guix.info,
it would be prudent to estimate the cost. CloudFront, like most Amazon
AWS services, is a "pay for what you use" model. The pricing is here:


To accurately estimate the cost, we need to know how many requests we
expect to receive, and how many bytes we expect to transfer out, during
a single month. Do we have information like this for berlin today?

Although I don't doubt that a CDN will perform better than what we have
now, I do think it would be good to measure the performance so that we
know for sure the money spent is actually providing a benefit. It would
be nice to have some data before and after to measure how availability
and performance have changed. Apart from anecdotes, what data do we
have to determine whether performance has improved after introducing a
CDN? For example, the following information could be useful:

* Network load on the origin server(s)
* Clients' latency to (the addresses pointed to by) ci.guix.info
* Clients' throughput while downloading substitutes from ci.guix.info

We don't log or collect client metrics, and that's fine. It could be
useful to add code to Guix to measure things like this when the user
asks to do so, but perhaps it isn't necessary. It may be good enough if
people just volunteer to manually gather some information and share it.
For example, you can define a shell function like this:

Toggle snippet (22 lines)
measure_get () {
curl -L \
-o /dev/null \
-w "url_effective: %{url_effective}\\n\
http_code: %{http_code}\\n\
num_connects: %{num_connects}\\n\
num_redirects: %{num_redirects}\\n\
remote_ip: %{remote_ip}\\n\
remote_port: %{remote_port}\\n\
size_download: %{size_download} B\\n\
speed_download: %{speed_download} B/s\\n\
time_appconnect: %{time_appconnect} s\\n\
time_connect: %{time_connect} s\\n\
time_namelookup: %{time_namelookup} s\\n\
time_pretransfer: %{time_pretransfer} s\\n\
time_redirect: %{time_redirect} s\\n\
time_starttransfer: %{time_starttransfer} s\\n\
time_total: %{time_total} s\\n" \
"$1"
}

See "man curl" for the meaning of each metric.

You can then use this function to measure a substitute download. Here's
an example in which I download a large substitute (linux-libre) from one
of my machines in Seattle:

Toggle snippet (19 lines)
$ measure_get https://berlin.guixsd.org/nar/gzip/1bq783rbkzv9z9zdhivbvfzhsz2s5yac-linux-libre-4.19 2>/dev/null
url_effective: https://berlin.guixsd.org/nar/gzip/1bq783rbkzv9z9zdhivbvfzhsz2s5yac-linux-libre-4.19
http_code: 200
num_connects: 1
num_redirects: 0
remote_ip: 141.80.181.40
remote_port: 443
size_download: 69899433 B
speed_download: 4945831.000 B/s
time_appconnect: 0.885277 s
time_connect: 0.459667 s
time_namelookup: 0.254210 s
time_pretransfer: 0.885478 s
time_redirect: 0.000000 s
time_starttransfer: 1.273994 s
time_total: 14.133584 s
$

Here, it took 0.459667 - 0.254210 = 0.205457 seconds (about 205 ms) to
establish the TCP connection after the DNS lookup. The average
throughput was 1924285 bytes per second (about 40 megabits per second,
where 1 megabit = 10^6 bits). It seems my connection to berlin is
already pretty good!

We can get more information about latency by using a tool like mtr:

Toggle snippet (7 lines)
$ sudo mtr -c 10 --report-wide --tcp -P 443 berlin.guixsd.org
Start: 2018-12-08T16:57:40-0800
HOST: localhost.localdomain Loss% Snt Last Avg Best Wrst StDev
[... I've omitted the intermediate hops because they aren't relevant ...]
13.|-- 141.80.181.40 0.0% 10 205.0 201.9 194.9 212.8 5.6

My machine's latency to berlin is about 202 ms, which matches what we
calculated above.

For experimentation, I've set up a CloudFront distribution at
berlin-mirror.marusich.info that uses berlin.guixsd.org as its origin
server. Let's repeat these steps to measure the performance of the
distribution from my machine's perspective (before I did this, I made
sure the GET would result in a cache hit by downloading the substitute
once before and verifying that the same remote IP address was used):

Toggle snippet (23 lines)
$ measure_get https://berlin-mirror.marusich.info/nar/gzip/1bq783rbkzv9z9zdhivbvfzhsz2s5yac-linux-libre-4.19 2>/dev/null
url_effective: https://berlin-mirror.marusich.info/nar/gzip/1bq783rbkzv9z9zdhivbvfzhsz2s5yac-linux-libre-4.19
http_code: 200
num_connects: 1
num_redirects: 0
remote_ip: 13.32.254.57
remote_port: 443
size_download: 69899433 B
speed_download: 9821474.000 B/s
time_appconnect: 0.607593 s
time_connect: 0.532417 s
time_namelookup: 0.511086 s
time_pretransfer: 0.608029 s
time_redirect: 0.000000 s
time_starttransfer: 0.663578 s
time_total: 7.117266 s
$ sudo mtr -c 10 --report-wide --tcp -P 443 berlin-mirror.marusich.info
Start: 2018-12-08T17:04:48-0800
HOST: localhost.localdomain Loss% Snt Last Avg Best Wrst StDev
[... I've omitted the intermediate hops because they aren't relevant ...]
14.|-- server-52-84-21-199.sea32.r.cloudfront.net 0.0% 10 19.8 20.3 14.3 28.9 4.9

Establishing the TCP connection took about 21 ms (which matches the mtr
output), and the throughput was about 79 megabits per second. (On this
machine, 100 Mbps is the current link speed, according to dmesg output.)
This means that in my case, when using CloudFront the latency is 10x
lower, and the throughput (for a cache hit) is 2x higher, than using
berlin.guixsd.org directly!

It would be interesting to see what the performance is for others.

Ricardo Wurmus <rekado@elephly.net> writes:

Toggle quote (4 lines)
> Large ISPs also provide CDN services. I already contacted Deutsche
> Telekom so that we can compare their CDN offer with the Amazon Cloudfont
> setup that Chris has configured.

That's great! There are many CDN services out there. I am unfamiliar
with most of them. It will be good to see how Deutsche Telekom's
offering compares to CloudFront.

FYI, CloudFront has edge locations in the following parts of the world:


Hartmut Goebel <hartmut@goebel-consult.de> writes:

Toggle quote (13 lines)
> Am 03.12.2018 um 17:12 schrieb Ludovic Courtès:
>> Thus, I’m thinking about using a similar setup, but hosting the mirror
>> on some Big Corp CDN or similar.
>
> Isn't this a contradiction: Building a free infrastructure relaying on
> servers from some Big Corporation? Let allow the privacy concerns
> raising when delivering data via some Big Corporation.
>
> If delivering "packages" works via static data without requiring any
> additional service, we could ask universities to host Guix, too. IMHO
> this is a much preferred solution since this is a decentralized publish
> infrastructure already in place for many GNU/Linux distributions.

I understand your concern about using a third-party service. However,
we wouldn't be using a CDN as a "software substitute", which is one of
the primary risks of using a web service today:


Instead, we would be using a CDN as a performance optimization that is
transparent to a Guix user. You seem unsettled by the idea of
entrusting any part of substitute delivery to a third party, but
concretely what risks do you foresee?

Regarding your suggestion to ask universities to host mirrors (really,
caching proxies), I think it could be a good idea. As Leo mentioned,
the configuration to set up an NGINX caching proxy of Hydra (or berlin)
is freely available in maintenance.git. Do you think we could convince
some universities to host caching proxies that just run an NGINX web
server using those configurations?

If we can accomplish that, it may still be helpful. If there is
interest in going down this path, I can explore some possibilities in
the Seattle area. If the university-owned caching proxies are easily
discoverable (i.e., we list them on the website), then users might
manually set their substitute URL to point to one that's close by.

Going further, if our DNS provider supports something like "geolocation
routing" for DNS queries, we might even be able to create DNS records
for ci.guix.info that point to those universities' caching proxies. In
this way, when a user resolves ci.guix.info, they would get the address
of a university-owned caching proxy close by. This could have the
benefits of requiring less money than a full-fledged CDN like Amazon
CloudFront, and also decentralizing the substitute delivery, while still
remaining transparent to Guix users. However, it would still require us
to rely on a third-party DNS service.

For example, Amazon Route 53 provides this sort of geolocation routing:


I wouldn't be surprised if there are other DNS providers out there who
offer something similar. However, I also wouldn't be surprised if the
overall performance of CloudFront turns out to be better.

"Thompson, David" <dthompson2@worcester.edu> writes:

Toggle quote (11 lines)
> If AWS CloudFront is the path chosen, it may be worthwhile to follow
> the "infrastructure as code" practice and use CloudFormation to
> provision the CloudFront distribution and any other supporting
> resources. The benefit is that there would be a record of exactly
> *how* the project is using these commercial services and the setup
> could be easily reproduced. The timing is interesting here because I
> just attended the annual AWS conference on behalf of my employer and
> while I was there I felt inspired to write a Guile API for building
> CloudFormation "stacks". You can see a small sample of what it does
> here: https://gist.github.com/davexunit/db4b9d3e67902216fbdbc66cd9c6413e

Nice! That seems useful. I will have to play with it. I created my
distributions manually using the AWS Management Console, since it's
relatively easy to do. I agree it would be better to practice
"infrastructure as code."

On that topic, I've also heard good things about Terraform by HashiCorp,
which is available under the Mozilla Public License 2.0:


Here is a comparison of Terraform and CloudFormation:


I looked briefly into packaging Terraform for Guix. It's written in Go.
It seems possible, but I haven't invested enough time yet.

As a final option, since the AWS CLI is already packaged in Guix, we
could just drive CloudFormation or CloudFront directly from the CLI.

Meiyo Peng <meiyo.peng@gmail.com> writes:

Toggle quote (8 lines)
> I like the idea of IPFS. We should try it. It would be great if it works
> well.
>
> If at some point we need to setup traditional mirrors like other major
> Gnu/Linux distros, I can contact my friends in China to setup mirrors in
> several universities. I was a member of LUG@USTC, which provides the
> largest FLOSS mirror in China.

IPFS would be neat. So would Gnunet. Heck, even a publication
mechanism using good old BitTorrent would be nice. All of these would
require changes to Guix, I suppose.

A CDN would require no changes to Guix, and that's part of why it's so
appealing.

--
Chris
-----BEGIN PGP SIGNATURE-----
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=H3rW
-----END PGP SIGNATURE-----

L
L
Ludovic Courtès wrote on 9 Dec 2018 14:58
(name . Hartmut Goebel)(address . hartmut@goebel-consult.de)
87pnua244k.fsf@gnu.org
Hi Hartmut,

Hartmut Goebel <hartmut@goebel-consult.de> skribis:

Toggle quote (14 lines)
> Am 09.12.2018 um 04:33 schrieb Chris Marusich:
>> Instead, we would be using a CDN as a performance optimization that is
>> transparent to a Guix user. You seem unsettled by the idea of
>> entrusting any part of substitute delivery to a third party, but
>> concretely what risks do you foresee?
>
> I have serious privacy concerns.
>
> TL;DR: A CDN is a centralized infrastructure, allowing to collect
> information about valuable vulnerability information of almost all
> Guix-users and -systems. This is might become a thread to freedom of
> speech, human rights, democracy and economics. Guix should build on a
> decentralized infrastructure.

Heck it would be ironic to find myself arguing in favor of centralized
commercial services. So I won’t do that. :-)

Clearly, I do understand the concerns you list. As a maintainer, I’m
looking for solutions that can address real problems (availability of
substitutes and bandwidth) while not being a threat to our user’s
privacy and security.

The operator of a substitute server (or caching proxy), in general,
knows which IPs downloaded vulnerable software. This is the main
threat.

This can be mitigated by talking to nearby mirrors and not just
ci.guix.info, a feature we implemented a year ago (see
or by using several substitute servers, or by not using (or not always
using) substitutes. Few distros have all these options.

We might also be able to somehow balance requests between several CDNs
or mirrors.

But again, medium- to long-term, the goal is to move towards IPFS or
GNUnet/Bittorrent. IPFS is attractive because it would probably require
no modifications to ‘guix substitutes’ and only minor changes to ‘guix
publish’ since the IPFS daemon has an HTTP interface.

Toggle quote (17 lines)
>> Regarding your suggestion to ask universities to host mirrors (really,
>> caching proxies), I think it could be a good idea. As Leo mentioned,
>> the configuration to set up an NGINX caching proxy of Hydra (or berlin)
>> is freely available in maintenance.git. Do you think we could convince
>> some universities to host caching proxies that just run an NGINX web
>> server using those configurations?
>
> The difference is: For a traditional "ftp"-mirror, an organization just
> needs to add another source to its existing configuration and administer
> to the save way as all other mirrors. Whereas for a caching proxy they
> need to change the setup of the web-server and learn how to administer
> the cache. This difference might make it difficult to convince
> organizations to mirror.
>
> I could try and ask a few organizations in my area, but I would need
> figures for this.

What would you need to know? ‘guix weather’ can provide info about
storage size.

Thanks,
Ludo’.
L
L
Ludovic Courtès wrote on 9 Dec 2018 16:59
CDN performance
(name . Chris Marusich)(address . cmmarusich@gmail.com)
871s6qzo6m.fsf_-_@gnu.org
Hello Chris,

Chris Marusich <cmmarusich@gmail.com> skribis:

Toggle quote (6 lines)
> Regarding DNS, it would be nice if we could use an official GNU
> subdomain. If we can't use a GNU subdomain, we should at least make
> sure we have some kind of DNS auto-renewal set up so that nobody can
> poach our domain names. And the operators should take appropriate
> precautions when sharing any credentials used for managing it all.

Agreed.

Regarding the GNU sub-domain, as I replied to Meiyo, I’m in favor of it,
all we need is someone to champion setting it up.

Toggle quote (6 lines)
> Regarding CDNs, I definitely think it's worth a try! Even Debian is
> using CloudFront (cloudfront.debian.net). In fact, email correspondence
> suggests that as of 2013, Amazon may even have been paying for it!
>
> https://lists.debian.org/debian-cloud/2013/05/msg00071.html

(Note that debian.net is not Debian, and “there’s no cloud, only other
people’s computer” as the FSFE puts it.)

Toggle quote (12 lines)
> Although I don't doubt that a CDN will perform better than what we have
> now, I do think it would be good to measure the performance so that we
> know for sure the money spent is actually providing a benefit. It would
> be nice to have some data before and after to measure how availability
> and performance have changed. Apart from anecdotes, what data do we
> have to determine whether performance has improved after introducing a
> CDN? For example, the following information could be useful:
>
> * Network load on the origin server(s)
> * Clients' latency to (the addresses pointed to by) ci.guix.info
> * Clients' throughput while downloading substitutes from ci.guix.info

Note that performance is one aspect; another one is availability (we’ve
seen that with the recent hydra.gnu.org outage!). We know we’ll always
win in terms of availability by having a CDN in front of our servers.

That said, measuring performance is very useful and it’s great that we
can benefit from your expertise here!

Toggle quote (6 lines)
> Here, it took 0.459667 - 0.254210 = 0.205457 seconds (about 205 ms) to
> establish the TCP connection after the DNS lookup. The average
> throughput was 1924285 bytes per second (about 40 megabits per second,
> where 1 megabit = 10^6 bits). It seems my connection to berlin is
> already pretty good!

Indeed. The bandwidth problem on berlin is when you’re the first to
download a nar and it’s not been cached by nginx yet. In that case, you
get very low bandwidth (like 10 times less than when the item is cached
by nginx.) I’ve looked into it, went as far as strace’ing nginx, but
couldn’t find the reason of this.

Do you any idea?

Toggle quote (7 lines)
> Establishing the TCP connection took about 21 ms (which matches the mtr
> output), and the throughput was about 79 megabits per second. (On this
> machine, 100 Mbps is the current link speed, according to dmesg output.)
> This means that in my case, when using CloudFront the latency is 10x
> lower, and the throughput (for a cache hit) is 2x higher, than using
> berlin.guixsd.org directly!

Impressive.

Toggle quote (2 lines)
> It would be interesting to see what the performance is for others.

I’ve tried this from home (in France, with FTTH):

Toggle snippet (34 lines)
$ measure_get https://berlin.guixsd.org/nar/gzip/1bq783rbkzv9z9zdhivbvfzhsz2s5yac-linux-libre-4.19 2>/dev/null
url_effective: https://berlin.guixsd.org/nar/gzip/1bq783rbkzv9z9zdhivbvfzhsz2s5yac-linux-libre-4.19
http_code: 200
num_connects: 1
num_redirects: 0
remote_ip: 141.80.181.40
remote_port: 443
size_download: 69899433 B
speed_download: 1522001.000 B/s
time_appconnect: 0.178892 s
time_connect: 0.049649 s
time_namelookup: 0.000422 s
time_pretransfer: 0.178934 s
time_redirect: 0.000000 s
time_starttransfer: 0.278312 s
time_total: 45.926021 s
$ measure_get https://berlin-mirror.marusich.info/nar/gzip/1bq783rbkzv9z9zdhivbvfzhsz2s5yac-linux-libre-4.19 2>/dev/null
url_effective: https://berlin-mirror.marusich.info/nar/gzip/1bq783rbkzv9z9zdhivbvfzhsz2s5yac-linux-libre-4.19
http_code: 200
num_connects: 1
num_redirects: 0
remote_ip: 2600:9000:2116:6e00:c:49d4:5a80:93a1
remote_port: 443
size_download: 69899433 B
speed_download: 20803402.000 B/s
time_appconnect: 0.552008 s
time_connect: 0.482477 s
time_namelookup: 0.467598 s
time_pretransfer: 0.552157 s
time_redirect: 0.000000 s
time_starttransfer: 0.735758 s
time_total: 3.360500 s

Wall-clock time is less than a tenth; woow.

Thanks for sharing your insight and scripts!

Ludo’.
H
H
Hartmut Goebel wrote on 9 Dec 2018 13:12
Re: Using a CDN or some other mirror?
d870d06a-a95c-2b0b-a196-b5166d50400a@goebel-consult.de
Am 09.12.2018 um 04:33 schrieb Chris Marusich:
Toggle quote (5 lines)
> Instead, we would be using a CDN as a performance optimization that is
> transparent to a Guix user. You seem unsettled by the idea of
> entrusting any part of substitute delivery to a third party, but
> concretely what risks do you foresee?

I have serious privacy concerns.

TL;DR: A CDN is a centralized infrastructure, allowing to collect
information about valuable vulnerability information of almost all
Guix-users and -systems. This is might become a thread to freedom of
speech, human rights, democracy and economics. Guix should build on a
decentralized infrastructure.

A distribution provider gets a notion which system is running which
software in which version. In case of guix, the provider even gets the
exact version of the software and all its dependencies. Combining this
with the rise of IPv6, which per default uses the MAC address as part of
the IP address, actually allows identifying a single system.

This information is extremely valuable for all kinds of attackers as it
eases attacking a system a lot. This becomes a thread to

* to opposition members, dissidents and human rights activists as the
intelligent agencies can target these persons much more precisely,
* to companies all over the world as many countries do industrial
espionage.

This becomes even worst when using a CDN, since the CDN is a centralized
system: A single CDN provider gains knowledge for almost all systems all
over the world. Which means: this valuable vulnerability information is
collected at a single place. Intelligence agencies might be keen on
getting access to this information and a centralized system makes it
easy for them. And there is evidence they actually collect this
information [*].

This gets even worse when the CDN belongs to one of these companies
compiling personal profiles, like Google, Facebook or Tencent. Amazon
belongs to this group.

I have the strong opinion that Guix should build on a decentralized
infrastructure to support keeping the freedom of speech, democracy and
human rights.

[*] Actually it is known the US-American intelligence agencies have
equipment placed at Verizon to collect all kind of data [1]. One can
reason the same is true for other big providers in the US. The USA has
the FISA act AFAIU enforcing US companies to collaborate in industrial
espionage. In Germany it is known that the BND is extracting high-volume
data at the central internet exchange (DE-CIX) [2]. One can reason such
also happens in other countries, esp. members of the five-eyes, France,
Russia, China, Israel, Saudi Arabia, Iran, Irak, etc.

Toggle quote (7 lines)
> Regarding your suggestion to ask universities to host mirrors (really,
> caching proxies), I think it could be a good idea. As Leo mentioned,
> the configuration to set up an NGINX caching proxy of Hydra (or berlin)
> is freely available in maintenance.git. Do you think we could convince
> some universities to host caching proxies that just run an NGINX web
> server using those configurations?

The difference is: For a traditional "ftp"-mirror, an organization just
needs to add another source to its existing configuration and administer
to the save way as all other mirrors. Whereas for a caching proxy they
need to change the setup of the web-server and learn how to administer
the cache. This difference might make it difficult to convince
organizations to mirror.

I could try and ask a few organizations in my area, but I would need
figures for this.


internet for e.g. "cia verizon espionage"
[2]
[4]

--
+++hartmut

| Hartmut Goebel | |
| hartmut@goebel-consult.de | www.goebel-consult.de |
Attachment: file
M
M
Meiyo Peng wrote on 11 Dec 2018 06:17
Re: CDN performance
(name . Ludovic Courtès)(address . ludo@gnu.org)
874lbk63se.fsf@gmail.com
Hi,

I have changed my mind. I don't care about the domain name of the
default substitute server any more. I am more worried about the network
speed within China. While I am writing this email, my computer is
downloading substitutes from ci.guix.info at an average speed of about
50kB/s.

After careful thought, I realized the new CDN won't benefit China
residents as planned. Any popular CDN outside China is significantly
throttled by ISP/GFW and the situation is worse every year. A CDN will
be a great improvement for western countries but not for many asia
countries.

I have been running a nginx caching mirror at home since the first time
I met guix. My first attempt to install guix took me one afternoon and
one night. Most of the time was wasted on downloading substitutes. After
that, I searched online and found the nginx mirror configuration in
guix/maintenance.git. Then I set up my own caching mirror. It would be
extremely painful for me to play with guix during these months without
this caching mirror. Debian was my favorite GNU/Linux distro before I
discovered Guix. It usually costs me only 30 minutes to install Debian
from scratch. Other GNU/Linux distros are about the same. If guix takes
more than 5 hours to install, most new users will be scared away.

Although now there are only a few Chinese guix users, I still prefer to
set up a caching mirror within China. The user experience improvement
will be significant. Like Hartmut Goebel described, I need information
about "expected storage demand, expected transfer traffic from upstream,
how often updates are published, how long to hold copies" so I can write
a proposal to maintainers of mirrors.ustc.edu.cn to add a new mirror for
guix. I tried the `guix weather` command, but it failed with an error:

#+BEGIN_EXAMPLE
computing 8,821 package derivations for x86_64-linux...
guix weather: error: lstat: No such file or directory: "/src/config/packages/patches/libjxr-add-cmakelists-file.patch"
#+END_EXAMPLE

Can anyone provide the numbers for me? The numbers of
mirror.hydra.gnu.org would be a good reference.

In addition, I think we should put up a page on guix's website providing
information about available substitute servers including address,
backend build farm, geolocation, maintainer, so users (many asia
residents) can pick the best substitute server for themselves. This will
be necessary until alternative solutions like IPFS or GNUnet are
implemented.

This email only describes the situation in China. I don't intend to
discourage you on the CDN plan. There is no conflicts between my
proposal to set up a caching mirror in China and our plan to set up a
new CDN.

--
Meiyo Peng
M
M
Meiyo Peng wrote on 11 Dec 2018 06:59
(name . Zephyr Waitzman)(address . i@wi24rd.ml)
87woog4n9i.fsf@gmail.com
Hi,
Zephyr Waitzman <i@wi24rd.ml> writes:
Toggle quote (1 lines)
> ???????????????????????
My caching mirror is at home and it has an ip address of
192.168.x.x. You may connect to it if you have magic power.
BTW. Nice to see another Chinese here. What's your experience
with the default substitute server?
--
Meiyo Peng
M
M
Meiyo Peng wrote on 11 Dec 2018 07:14
(name . Zephyr Waitzman)(address . i@wi24rd.ml)
87pnu84mjw.fsf@gmail.com
Hi,
Zephyr Waitzman <i@wi24rd.ml> writes:
Toggle quote (2 lines)
> ?????????????????????????????? substitude ?????????????????????????????
> ????????????
Thank you for your feedback. That means we do need to set up a caching
mirror within China.
--
Meiyo Peng
G
G
Giovanni Biscuolo wrote on 11 Dec 2018 17:38
Re: Using a CDN or some other mirror?
8736r4hvcs.fsf@roquette.mug.biscuolo.net
Hi all,

my two cents...

(I can't still help with a public cache, I hope soon...)

Ludovic Courtès <ludo@gnu.org> writes:

[...]

Toggle quote (6 lines)
>> TL;DR: A CDN is a centralized infrastructure, allowing to collect
>> information about valuable vulnerability information of almost all
>> Guix-users and -systems. This is might become a thread to freedom of
>> speech, human rights, democracy and economics. Guix should build on a
>> decentralized infrastructure.

I completely agree with you, decentralization is the solution

unfortunately the **only functioning** way is to avoid current Internet,
since it's broken (https://youbroketheinternet.org/);I see GuixSD as an
integral part of The Project Map™ https://youbroketheinternet.org/map

...but to fix the situation we need a substantial GNUnet(work) effect
and for that we _need_ GuixSD substitutes to be easily and quickly
downloaded (can we avoid this asking potential adopters to be patient or
to build?)

maybe we should divide this task in two steps:

1. distributed substitutes: caching servers hosted by a network of
friendly institutions and companies donated to GNU/GuixSD, with a
haproxy frontend for geolocated load-balancing [1]

2. decentralized substitutes: caching servers on IPFS or better (since
it allows complete anonimity) on GNUnet

Toggle quote (3 lines)
> Heck it would be ironic to find myself arguing in favor of centralized
> commercial services. So I won’t do that. :-)

I see no problems with commercial services, _unfortunately_ nowadays
this *almost* always means centralized silos, usually exploited for
global surveillance (since Internet is broken)

[...]

Toggle quote (4 lines)
> The operator of a substitute server (or caching proxy), in general,
> knows which IPs downloaded vulnerable software. This is the main
> threat.

on Internet, and on IPFS? (sorry for the ignorance)

on GNUNet filesharing can be completely anonymous, but the performace is
degraded (so we need a large network effect here)

Toggle quote (9 lines)
> This can be mitigated by talking to nearby mirrors and not just
> ci.guix.info, a feature we implemented a year ago (see
> <https://gnu.org/s/guix/blog/2017/reproducible-builds-a-status-update/>),
> or by using several substitute servers, or by not using (or not always
> using) substitutes. Few distros have all these options.
>
> We might also be able to somehow balance requests between several CDNs
> or mirrors.

did someone explored an haproxy (with geolocation) solution?

is there a wip-haproxy attempt?

[...]

HTH
Giovanni


[1] in the next few weeks I'm going to test an haproxy instance with
geolocated ACLs following this directions
https:/www.haproxy.com/blog/use-geoip-database-within-haproxy/

--
Giovanni Biscuolo

Xelera IT Infrastructures
-----BEGIN PGP SIGNATURE-----
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=WG9l
-----END PGP SIGNATURE-----

C
C
Chris Marusich wrote on 13 Dec 2018 08:11
Re: CDN performance
(name . Meiyo Peng)(address . meiyo.peng@gmail.com)
877egdyk82.fsf@gmail.com
Meiyo Peng <meiyo.peng@gmail.com> writes:

Toggle quote (6 lines)
> After careful thought, I realized the new CDN won't benefit China
> residents as planned. Any popular CDN outside China is significantly
> throttled by ISP/GFW and the situation is worse every year. A CDN will
> be a great improvement for western countries but not for many asia
> countries.

Could you try running the measure_get shell function I included in the
following email?


For convenience, here is the definition:

Toggle snippet (22 lines)
measure_get () {
curl -L \
-o /dev/null \
-w "url_effective: %{url_effective}\\n\
http_code: %{http_code}\\n\
num_connects: %{num_connects}\\n\
num_redirects: %{num_redirects}\\n\
remote_ip: %{remote_ip}\\n\
remote_port: %{remote_port}\\n\
size_download: %{size_download} B\\n\
speed_download: %{speed_download} B/s\\n\
time_appconnect: %{time_appconnect} s\\n\
time_connect: %{time_connect} s\\n\
time_namelookup: %{time_namelookup} s\\n\
time_pretransfer: %{time_pretransfer} s\\n\
time_redirect: %{time_redirect} s\\n\
time_starttransfer: %{time_starttransfer} s\\n\
time_total: %{time_total} s\\n" \
"$1"
}

Specifically, I am curious to know what performance you get when you run


from a computer in China. Please be sure to run it two times in a row,
to ensure that CloudFront has cached the object.

CloudFront has edge locations in Hong Kong, so I am curious to know what
performance improvement, if any, you observe.

--
Chris
-----BEGIN PGP SIGNATURE-----
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=h//j
-----END PGP SIGNATURE-----

C
C
Chris Marusich wrote on 13 Dec 2018 09:05
(name . Ludovic Courtès)(address . ludo@gnu.org)
87y38tx365.fsf@gmail.com
Ludovic Courtès <ludo@gnu.org> writes:

Toggle quote (3 lines)
> Regarding the GNU sub-domain, as I replied to Meiyo, I’m in favor of it,
> all we need is someone to champion setting it up.

I could help with this. Whom should I contact?

Toggle quote (9 lines)
>> Regarding CDNs, I definitely think it's worth a try! Even Debian is
>> using CloudFront (cloudfront.debian.net). In fact, email correspondence
>> suggests that as of 2013, Amazon may even have been paying for it!
>>
>> https://lists.debian.org/debian-cloud/2013/05/msg00071.html
>
> (Note that debian.net is not Debian, and “there’s no cloud, only other
> people’s computer” as the FSFE puts it.)

I do try to avoid the term "cloud" whenever possible. It's hard to
avoid when it's in the product name, though! A wise man once said, "A
cloud in the mind is an obstacle to clear thinking." ;-)

You may be right about debian.net. I don't know who owns that domain.
It's confusing, since debian.org is definitely owned by the Debian
project, and the following page says they're using Amazon CloudFront:


Maybe Debian still uses Amazon CloudFront, or maybe they don't any more.
In any case, I've found the following email thread, which documents a
thoughtful discussion regarding whether or not Debian should use a CDN.
They discussed many of the same concerns we're discussing here.


A summary, in the middle of the long thread, is here:


Later, part of the thread broke off and continued here:


That's as far as I've read.

Judging by that email thread, one of the reasons why Debian considered
using a CDN was because they felt that the cost, in terms of people
power, of maintaining their own "proto-CDN" infrastructure had grown too
great. I believe it! I think it would be ill-advised for the Guix
project to expend effort and capital on building and maintaining its own
CDN. I think it would be wiser to focus on developing a decentralized
substitute solution (GNUnet, IPFS, etc.).

That said, I still think that today Guix should provide a third-party
CDN option. For many Guix users, a CDN would improve performance and
availability of substitutes. Contracting with a third party to provide
the CDN service would require much less effort and capital than building
and maintaining a CDN from scratch. This would also enable the project
to focus more on building a decentralized substitute solution. And once
that decentralized solution is ready, it will be easy to just "turn off"
the CDN.

I also understand Hartmut's concerns. The risks he points out are
valid. Because of those risks, even if we make a third-party CDN option
available, some people will choose not to use it. For that reason, we
should not require Guix users to use a third-party CDN, just as we do
not require them to use substitutes from our build farm.

However, not everyone shares the same threat model. For example,
although some people choose not to trust substitutes from our build
farm, still others do. The choice is based on one's own individual
situation. Similarly, if we make a third-party CDN option available and
explain the risks of using it, Guix users will be able to make an
educated decision for themselves about whether or not to use it.

Toggle quote (14 lines)
>> Here, it took 0.459667 - 0.254210 = 0.205457 seconds (about 205 ms) to
>> establish the TCP connection after the DNS lookup. The average
>> throughput was 1924285 bytes per second (about 40 megabits per second,
>> where 1 megabit = 10^6 bits). It seems my connection to berlin is
>> already pretty good!
>
> Indeed. The bandwidth problem on berlin is when you’re the first to
> download a nar and it’s not been cached by nginx yet. In that case, you
> get very low bandwidth (like 10 times less than when the item is cached
> by nginx.) I’ve looked into it, went as far as strace’ing nginx, but
> couldn’t find the reason of this.
>
> Do you any idea?

I made a typo here. The value "1924285" should have been "4945831",
which is what measure_get printed. However, the intended result (40
Mbps) is still correct.

Actually, I thought 40 megabits per second was pretty great for a
single-threaded file transfer that originated in Europe (I think?) and
terminated in Seattle (via my residential Comcast downlink). I
requested that particular file many times before that final test run, so
it was probably already cached by nginx.

However, I believe you when you say that it's slow the first time you
download the substitute from berlin. What path does the data take from
its origin through berlin? If berlin needs to download the initial file
from another server, perhaps the connection between berlin and that
server is the bottleneck? Maybe we should discuss that in a different
email thread, though.

Toggle quote (6 lines)
> I’ve tried this from home (in France, with FTTH):
>
> [...]
>
> speed_download: 20803402.000 B/s

Wow, that's 166 megabits per second! I'm jealous. :-)

Toggle quote (2 lines)
> Wall-clock time is less than a tenth; woow.

I expect others will see similar performance improvements, as long as
they are close to the edge locations provided by Amazon CloudFront and
their local ISP downlink is fast enough to see a benefit. Again, the
edge locations are here:


Here's a direct link to their current map:


--
Chris
-----BEGIN PGP SIGNATURE-----
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=kuJA
-----END PGP SIGNATURE-----

G
G
Giovanni Biscuolo wrote on 13 Dec 2018 11:41
878t0thfp9.fsf@roquette.mug.biscuolo.net
Hi Chris,

nice to see this discussion, IMHO how GuixSD subsitutes are distributed
is a key issue in our ecosystem and is _all_ about privacy and metadata
*mass* collection

most "normal users" are not concerned about this so they are fine with
super-centralization since it's a convenience... not us :-)

personally I've come to GuixSD because I see this project as a key part
in liberating me from this class of problems

Chris Marusich <cmmarusich@gmail.com> writes:

[...]

Toggle quote (4 lines)
> A summary, in the middle of the long thread, is here:
>
> https://lists.debian.org/debian-project/2013/10/msg00074.html

thank you for the reference, I've only read this summary

the key part of it IMHO is
"Q: Do CDNs raise more security/privacy concerns than our mirrors?"

and the related subthread

the quick reply to the above question is: yes, CDNs raise more
secutiry/privacy concerns than "distributed mirrors"

obviuosly "distributed mirrors" _does_ rise some security/privacy
concerns but *centralization*... much more

[...]

Toggle quote (5 lines)
> Judging by that email thread, one of the reasons why Debian considered
> using a CDN was because they felt that the cost, in terms of people
> power, of maintaining their own "proto-CDN" infrastructure had grown too
> great.

I'm still new to guixsd but understood enough to se we are much more
well equipped to maintain our distributed network of substitutes caching
servers... **transparently** configured :-)

[...]

Toggle quote (4 lines)
> I also understand Hartmut's concerns. The risks he points out are
> valid. Because of those risks, even if we make a third-party CDN option
> available, some people will choose not to use it.

probably I'll be one of those, I'm considering to maintain a caching
substitute server in a "semi-trusted" colocated space and I'd be very
happy to share that server with the community

[...]

Toggle quote (4 lines)
> However, not everyone shares the same threat model. For example,
> although some people choose not to trust substitutes from our build
> farm, still others do.

for this very reason IMHO we should work towards a network of **very
trusted** build farms directly managed and controlled by the GuixSD
project sysadmins; if build farms will be able to quickly provide
substitutes, caching mirrors will be _much more_ effective than today

... and a network of "automated guix challenge" servers to spot
not-reproducible software in GuixSD

with a solid infrastructure of "scientifically" trustable build farms,
there are no reasons not to trust substitutes servers (this implies
working towards 100% reproducibility of GuixSD)

Toggle quote (5 lines)
> The choice is based on one's own individual
> situation. Similarly, if we make a third-party CDN option available and
> explain the risks of using it, Guix users will be able to make an
> educated decision for themselves about whether or not to use it.

that's an option... like a "last resort" in order to be able to use
guixSD :-)

we could also teach people how to setup their own caching servers and
possibly share them with the rest of the local community (possibly with
some coordination effort from the project sysadmins)

for Milan I've plans to setup such caching mirror in Jan 2019

[...]

happy hacking!
Gio

--
Giovanni Biscuolo

Xelera IT Infrastructures
-----BEGIN PGP SIGNATURE-----
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=jw02
-----END PGP SIGNATURE-----

H
H
Hartmut Goebel wrote on 14 Dec 2018 09:35
Re: Using a CDN or some other mirror?
(name . Ludovic Courtès)(address . ludo@gnu.org)
65fa6be4-3ddc-703b-708d-e223c2474d0b@crazy-compilers.com
Am 09.12.2018 um 14:58 schrieb Ludovic Courtès:
Toggle quote (5 lines)
>> I could try and ask a few organizations in my area, but I would need
>> figures for this.
> What would you need to know? ‘guix weather’ can provide info about
> storage size.

I don't know yet, which info the admins need for a decision. FMPOV I'd
says: Disk-space and traffic to be expected.

`guix weather` only provides the disk-space, but even this is not
obvious for me:

  13912.1 MiB of nars (compressed)
  41176.6 MiB on disk (uncompressed)

From reading the manual, I assume 13.9 GB are required on the server
(which is quite a lot IMHO). Is this correct?

--
+++hartmut

| Hartmut Goebel | |
| hartmut@goebel-consult.de | www.goebel-consult.de |
Attachment: file
P
P
Pierre Neidhardt wrote on 14 Dec 2018 10:02
(name . Hartmut Goebel)(address . h.goebel@crazy-compilers.com)
87tvjgeb12.fsf@ambrevar.xyz
Talking about this, I recently discussed with Ludovic the idea of compressing
nars in Lzip instead of gzip.


I can work on some Lzip guile-bindings, it should be quite easy, then we could
save some 10-50% (!!!) disk usage.

--
Pierre Neidhardt
-----BEGIN PGP SIGNATURE-----

iQEzBAEBCAAdFiEEUPM+LlsMPZAEJKvom9z0l6S7zH8FAlwTcakACgkQm9z0l6S7
zH/4QAf/Rn2lKzmd2wFCOMm2Dhr3e5Z507oAZv+6Ci0P7qxG51nW1A5rTQYr31qX
r8s++AUgD88nooX8ZtvKv13rkxWkdifnu4iXuwaRM/2FmXTp2NfQ8ZNvQi+/CKrc
xbDH8+1o6f0aID0c2EFSxo9eQn73COegujH+PZHHrgpBVHYe9e9ZC/54ACC2m9lA
S9u1Bj05TAJBpSxSbhe0wp/j47gn4ealuFHRqXuXRhy/fNMJB2uMS+ZcCL7wv+XY
1waYZDYElxrc13HsfCK3OfP8PiMplSTuq3lWYVvHysY/q0mXvTjola5Jdq9Z73KO
4BRXLSJg4qBesMnhLiHQVOrdJrAWhQ==
=zi5X
-----END PGP SIGNATURE-----

L
L
Ludovic Courtès wrote on 14 Dec 2018 11:26
guix.gnu.org sub-domain
(name . Chris Marusich)(address . cmmarusich@gmail.com)
874lbg76b2.fsf_-_@gnu.org
Hi Chris,

Chris Marusich <cmmarusich@gmail.com> skribis:

Toggle quote (7 lines)
> Ludovic Courtès <ludo@gnu.org> writes:
>
>> Regarding the GNU sub-domain, as I replied to Meiyo, I’m in favor of it,
>> all we need is someone to champion setting it up.
>
> I could help with this. Whom should I contact?

We discussed this over the last few days in Paris and Julien (roptat on
IRC) volunteered to come up with a Knot service setup for bayfront.scm.
When that’s ready, we can contact the FSF sysadmins so they delegate to
bayfront.

I’m sure Julien wouldn’t mind getting some help or insight, so please do
get in touch!

Ludo’.
L
L
Ludovic Courtès wrote on 14 Dec 2018 11:35
Re: CDN performance
(name . Chris Marusich)(address . cmmarusich@gmail.com)
87pnu45rba.fsf@gnu.org
Hello!

Chris Marusich <cmmarusich@gmail.com> skribis:

Toggle quote (3 lines)
> Ludovic Courtès <ludo@gnu.org> writes:


[...]

Toggle quote (30 lines)
> Judging by that email thread, one of the reasons why Debian considered
> using a CDN was because they felt that the cost, in terms of people
> power, of maintaining their own "proto-CDN" infrastructure had grown too
> great. I believe it! I think it would be ill-advised for the Guix
> project to expend effort and capital on building and maintaining its own
> CDN. I think it would be wiser to focus on developing a decentralized
> substitute solution (GNUnet, IPFS, etc.).
>
> That said, I still think that today Guix should provide a third-party
> CDN option. For many Guix users, a CDN would improve performance and
> availability of substitutes. Contracting with a third party to provide
> the CDN service would require much less effort and capital than building
> and maintaining a CDN from scratch. This would also enable the project
> to focus more on building a decentralized substitute solution. And once
> that decentralized solution is ready, it will be easy to just "turn off"
> the CDN.
>
> I also understand Hartmut's concerns. The risks he points out are
> valid. Because of those risks, even if we make a third-party CDN option
> available, some people will choose not to use it. For that reason, we
> should not require Guix users to use a third-party CDN, just as we do
> not require them to use substitutes from our build farm.
>
> However, not everyone shares the same threat model. For example,
> although some people choose not to trust substitutes from our build
> farm, still others do. The choice is based on one's own individual
> situation. Similarly, if we make a third-party CDN option available and
> explain the risks of using it, Guix users will be able to make an
> educated decision for themselves about whether or not to use it.

That summarizes the situation very well.

In Paris, there was consensus on two things: that a CDN would be helpful
for performance and availability (though availability could suddenly be
worse if we run… out of money), and that it raises privacy concerns.

Based on this, some suggested having a way to opt out of the CDN
distribution: for example, we might make ci.guix.info point to the CDN
and also advertise berlin.guixsd.org as a way to bypass it.

Of course in parallel we want to do our best to develop decentralized
solutions.

Toggle quote (7 lines)
> However, I believe you when you say that it's slow the first time you
> download the substitute from berlin. What path does the data take from
> its origin through berlin? If berlin needs to download the initial file
> from another server, perhaps the connection between berlin and that
> server is the bottleneck? Maybe we should discuss that in a different
> email thread, though.

On the first hit, nginx fetches the file from ‘guix publish’, which is
also on berlin. ‘guix publish’ gets the file from local storage and
passes it with sendfile(2), so it’s as fast as can be.

Thanks for the info!

Ludo’.
L
L
Ludovic Courtès wrote on 14 Dec 2018 15:45
Re: Using a CDN or some other mirror?
(name . Hartmut Goebel)(address . h.goebel@crazy-compilers.com)
87a7l82mmg.fsf@gnu.org
Hello,

Hartmut Goebel <h.goebel@crazy-compilers.com> skribis:

Toggle quote (18 lines)
> Am 09.12.2018 um 14:58 schrieb Ludovic Courtès:
>>> I could try and ask a few organizations in my area, but I would need
>>> figures for this.
>> What would you need to know? ‘guix weather’ can provide info about
>> storage size.
>
> I don't know yet, which info the admins need for a decision. FMPOV I'd
> says: Disk-space and traffic to be expected.
>
> `guix weather` only provides the disk-space, but even this is not
> obvious for me:
>
>   13912.1 MiB of nars (compressed)
>   41176.6 MiB on disk (uncompressed)
>
> From reading the manual, I assume 13.9 GB are required on the server
> (which is quite a lot IMHO). Is this correct?

If you’re running a caching proxy, you’ll need 13G. However note that
it’s only for one architecture and one revision of Guix. The total
space needed is obviously a function of time (number of Guix revisions
served) and number of architectures.

You could choose an expiration time in your caching proxy that satisfies
your disk space constraints, though.

Our machines that run Cuirass + ‘guix publish --cache’ need roughly
41+13G since they contain both /gnu/store and /var/cache/guix/publish.

HTH!

Ludo’.
L
L
Ludovic Courtès wrote on 14 Dec 2018 15:48
Compressing nars with lzip or similar
(name . Pierre Neidhardt)(address . mail@ambrevar.xyz)
874lbg2mh3.fsf_-_@gnu.org
Pierre Neidhardt <mail@ambrevar.xyz> skribis:

Toggle quote (8 lines)
> Talking about this, I recently discussed with Ludovic the idea of compressing
> nars in Lzip instead of gzip.
>
> http://lzip.nongnu.org/lzip.html (benchmark included)
>
> I can work on some Lzip guile-bindings, it should be quite easy, then we could
> save some 10-50% (!!!) disk usage.

That would be sweet!

Though we have to keep in mind that today’s ‘guix substitute’ doesn’t
know about lzip at all (see (guix scripts substitute) and
‘call-with-decompressed-port’.) So while lzip would be my preference,
we may not be able to use it until we can be sure our users can actually
unpack it.

Ludo’.
P
P
Pierre Neidhardt wrote on 14 Dec 2018 16:21
(name . Ludovic Courtès)(address . ludo@gnu.org)
878t0sdthe.fsf@ambrevar.xyz
Absolutely, but once the bindings are up and running, it should be
straightforward to update (guix scripts substitute), no?

--
Pierre Neidhardt
-----BEGIN PGP SIGNATURE-----

iQEzBAEBCAAdFiEEUPM+LlsMPZAEJKvom9z0l6S7zH8FAlwTyn0ACgkQm9z0l6S7
zH+zPAf7BkIiDgq7m3duDpfVf3z4wjB9W+9O6+8N7ilk6GEarHzKN7Ed5yfxiKBK
8f1/eZPhAQusAmOlucIa5cwed6OgdJg7saYUvwMfG+yDRh2SBRaNr79/ISjRfrm7
JqdHFABRZQVUeS4kMjeIcUNH7um1ZICgetV9X3Jb4zZUgq8FiEK39NGSQIeUSn7A
cbcLYK5fCzXyHikhiDmbWB/k9Xs7ej8poOFYH0wjLSb6wcYCCuuoEXY8qef+zFHQ
RvL6t7Y1w5eO6dfxYPHbcgUbsDdSASVpbYsR5J6QPjve1zfJ+7YX0edXSbz1TnZ6
8LLMlMC3nuznI8l9MpYt2V15zPkp/w==
=fPIo
-----END PGP SIGNATURE-----

M
M
Mark H Weaver wrote on 15 Dec 2018 02:40
Re: CDN performance
(name . Giovanni Biscuolo)(address . g@xelera.eu)
874lbfd0sd.fsf@netris.org
Hi Giovanni,

Giovanni Biscuolo <g@xelera.eu> writes:
Toggle quote (4 lines)
> with a solid infrastructure of "scientifically" trustable build farms,
> there are no reasons not to trust substitutes servers (this implies
> working towards 100% reproducibility of GuixSD)

What does "scientifically trustable" mean?

Mark
P
P
Pierre Neidhardt wrote on 15 Dec 2018 13:17
Re: Compressing nars with lzip or similar
(name . Ludovic Courtès)(address . ludo@gnu.org)
87k1kbrnlq.fsf@ambrevar.xyz
I've done some quick research over the various options.

- lzip: better than gz and bzip2 for sure, possibly better than xz (at
least according to the author).

- plzip: for "parallel lzip". With 4 threads I was able to compress
icecat 2.5x faster. It used 5x more memory though. The compression
ratio is 1-2% worse.

- lrzip: it would crash whenever I would change the compression level.
Seems less stable. It's as fast as plzip, while being 1-2% less
compressed. I don't think it's worth using.

All in all, lzip is a definite win over most options. The main question
is: lzip or plzip?

In my opinion it makes more sense to use plzip, gotta put those cores to
some use (on the user side at least)! Not sure what the build farm
would think about this though.

Finally, plzip can be quite memory-intensive. When passed "--threads=1"
however, it gets closer to lzip (bit faster, bit more memory). So maybe
the ideal would be to support plzip with a user-settable "threads" option.

Thoughts?

--
Pierre Neidhardt
-----BEGIN PGP SIGNATURE-----

iQEzBAEBCAAdFiEEUPM+LlsMPZAEJKvom9z0l6S7zH8FAlwU8MEACgkQm9z0l6S7
zH+5owgAhdRMia1xxIiCANu8UorM980EiPuGsTAO/VnnLpleGnQsdfQKjU7SWsk+
3aV9B/zHmyyouA902HTQK0h5HxzXfW2QivrZx7RtkOFXnckbvMjt7epT0UzaaK+v
E+zjC5o16DmDiRyYdmqttDnSNL4Kp30zawRDwN+IXYaWm3L1cb6qSU7H4UgmbFng
80MHofqdU/2Uk042pNwZNQUZPjtIWSX611LJGT7YmrfscyDDBhijXTIjiNaIKa6b
Tm7Vr7HLN100k3gzPVEBONfrDFMBfOqNHfxy8Y1+79c0eMFqHbvSDKpghxRCrByk
lhGLOPPKrvH3LWkJHE2MHh/z8pIb5Q==
=gxzl
-----END PGP SIGNATURE-----

L
L
Ludovic Courtès wrote on 15 Dec 2018 19:04
(name . Pierre Neidhardt)(address . mail@ambrevar.xyz)
87tvjey8co.fsf@gnu.org
Pierre Neidhardt <mail@ambrevar.xyz> skribis:

Toggle quote (3 lines)
> Absolutely, but once the bindings are up and running, it should be
> straightforward to update (guix scripts substitute), no?

Sure! But it will take time before we can assume all our users are
running it.

Ludo’.
L
L
Ludovic Courtès wrote on 15 Dec 2018 19:06
(name . Pierre Neidhardt)(address . mail@ambrevar.xyz)
87o99my89u.fsf@gnu.org
Pierre Neidhardt <mail@ambrevar.xyz> skribis:

Toggle quote (16 lines)
> I've done some quick research over the various options.
>
> - lzip: better than gz and bzip2 for sure, possibly better than xz (at
> least according to the author).
>
> - plzip: for "parallel lzip". With 4 threads I was able to compress
> icecat 2.5x faster. It used 5x more memory though. The compression
> ratio is 1-2% worse.
>
> - lrzip: it would crash whenever I would change the compression level.
> Seems less stable. It's as fast as plzip, while being 1-2% less
> compressed. I don't think it's worth using.
>
> All in all, lzip is a definite win over most options. The main question
> is: lzip or plzip?

‘guix publish’ has its own worker pool and handles parallelism
internally, so in that context plain sequential lzip would be more
appropriate IMO.

Ludo’.
C
C
Chris Marusich wrote on 16 Dec 2018 00:20
Re: guix.gnu.org sub-domain
(name . Ludovic Courtès)(address . ludo@gnu.org)
87va3us7gf.fsf@gmail.com
Hi Ludo,

Ludovic Courtès <ludo@gnu.org> writes:

Toggle quote (3 lines)
> I’m sure Julien wouldn’t mind getting some help or insight, so please do
> get in touch!

OK, I'll speak privately with Julien about the DNS setup to avoid adding
noise to this email thread.

--
Chris
-----BEGIN PGP SIGNATURE-----
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=Iuh7
-----END PGP SIGNATURE-----

M
M
Meiyo Peng wrote on 17 Dec 2018 07:48
Re: CDN performance
(name . Chris Marusich)(address . cmmarusich@gmail.com)
87d0q01wdx.fsf@gmail.com
Hi Chris,

Chris Marusich <cmmarusich@gmail.com> writes:

Toggle quote (46 lines)
> Meiyo Peng <meiyo.peng@gmail.com> writes:
>
>> After careful thought, I realized the new CDN won't benefit China
>> residents as planned. Any popular CDN outside China is significantly
>> throttled by ISP/GFW and the situation is worse every year. A CDN will
>> be a great improvement for western countries but not for many asia
>> countries.
>
> Could you try running the measure_get shell function I included in the
> following email?
>
> https://lists.gnu.org/archive/html/guix-devel/2018-12/msg00192.html
>
> For convenience, here is the definition:
>
> measure_get () {
> curl -L \
> -o /dev/null \
> -w "url_effective: %{url_effective}\\n\
> http_code: %{http_code}\\n\
> num_connects: %{num_connects}\\n\
> num_redirects: %{num_redirects}\\n\
> remote_ip: %{remote_ip}\\n\
> remote_port: %{remote_port}\\n\
> size_download: %{size_download} B\\n\
> speed_download: %{speed_download} B/s\\n\
> time_appconnect: %{time_appconnect} s\\n\
> time_connect: %{time_connect} s\\n\
> time_namelookup: %{time_namelookup} s\\n\
> time_pretransfer: %{time_pretransfer} s\\n\
> time_redirect: %{time_redirect} s\\n\
> time_starttransfer: %{time_starttransfer} s\\n\
> time_total: %{time_total} s\\n" \
> "$1"
> }
>
> Specifically, I am curious to know what performance you get when you run
>
> measure_get https://berlin-mirror.marusich.info/nar/gzip/1bq783rbkzv9z9zdhivbvfzhsz2s5yac-linux-libre-4.19
>
> from a computer in China. Please be sure to run it two times in a row,
> to ensure that CloudFront has cached the object.
>
> CloudFront has edge locations in Hong Kong, so I am curious to know what
> performance improvement, if any, you observe.

Sorry for the delay. My computer was reinstalled with Windows and taken
away by my girlfriend. So I have been waiting for my new computer that I
bought online to arrive. Finally, it arrived yesterday and I
successfully installed Guix on it.

I tested your script several times.

1. Tested today at home. China Unicom home broadband. 50Mb/s.

The result is slow as usual. curl failed once.
berlin-mirror.marusich.info is resolved to Seattle, WA, US.

#+BEGIN_EXAMPLE
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
55 66.6M 55 36.9M 0 0 17926 0 1:04:59 0:36:02 0:28:57 17733
http_code: 200
num_connects: 1
num_redirects: 0
remote_ip: 52.85.158.151
remote_port: 443
size_download: 38764357 B
speed_download: 17926.000 B/s
time_appconnect: 6.078850 s
time_connect: 3.006821 s
time_namelookup: 2.659785 s
time_pretransfer: 6.079097 s
time_redirect: 0.000000 s
time_starttransfer: 9.626001 s
time_total: 2162.379211 s
curl: (92) HTTP/2 stream 0 was not closed cleanly: INTERNAL_ERROR (err 2)

% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
100 66.6M 100 66.6M 0 0 109k 0 0:10:25 0:10:25 --:--:-- 241k
http_code: 200
num_connects: 1
num_redirects: 0
remote_ip: 52.85.158.22
remote_port: 443
size_download: 69899433 B
speed_download: 111816.000 B/s
time_appconnect: 3.507528 s
time_connect: 2.650373 s
time_namelookup: 2.261801 s
time_pretransfer: 3.507637 s
time_redirect: 0.000000 s
time_starttransfer: 5.995298 s
time_total: 625.129571 s

% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
100 66.6M 100 66.6M 0 0 109k 0 0:10:23 0:10:23 --:--:-- 141k
http_code: 200
num_connects: 1
num_redirects: 0
remote_ip: 52.85.158.22
remote_port: 443
size_download: 69899433 B
speed_download: 112187.000 B/s
time_appconnect: 2.280972 s
time_connect: 1.407197 s
time_namelookup: 1.056180 s
time_pretransfer: 2.281234 s
time_redirect: 0.000000 s
time_starttransfer: 3.167703 s
time_total: 623.061584 s
#+END_EXAMPLE


2. Tested 3 days ago at my office. China Telecom enterprise broadband. 50Mb/s.

Unusually fast! berlin-mirror.marusich.info is resolved to Seattle, WA,
US. I have no idea why it's so fast that day.

#+BEGIN_EXAMPLE
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
100 66.6M 100 66.6M 0 0 1364k 0 0:00:50 0:00:50 --:--:-- 1352k
http_code: 200
num_connects: 1
num_redirects: 0
remote_ip: 13.35.20.109
remote_port: 443
size_download: 69899433 B
speed_download: 1397429.000 B/s
time_appconnect: 2.432387 s
time_connect: 0.200842 s
time_namelookup: 0.000446 s
time_pretransfer: 2.432659 s
time_redirect: 0.000000 s
time_starttransfer: 2.673045 s
time_total: 50.020945 s

% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
100 66.6M 100 66.6M 0 0 1592k 0 0:00:42 0:00:42 --:--:-- 2506k
http_code: 200
num_connects: 1
num_redirects: 0
remote_ip: 13.35.20.109
remote_port: 443
size_download: 69899433 B
speed_download: 1630687.000 B/s
time_appconnect: 0.653270 s
time_connect: 0.209455 s
time_namelookup: 0.001582 s
time_pretransfer: 0.658399 s
time_redirect: 0.000000 s
time_starttransfer: 0.883126 s
time_total: 42.865868 s
#+END_EXAMPLE


3. Tested today at my office. China Telecom enterprise broadband. 50Mb/s.

Slow as usual. berlin-mirror.marusich.info is still resolved to Seattle,
WA, US. I killed the program several times because it hung there with no
data transfer for a few minutes. The TCP connection was probably closed
by GFW. This is very common here.

#+BEGIN_EXAMPLE
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
100 66.6M 100 66.6M 0 0 48110 0 0:24:12 0:24:12 --:--:-- 41808
http_code: 200
num_connects: 1
num_redirects: 0
remote_ip: 52.85.158.151
remote_port: 443
size_download: 69899433 B
speed_download: 48110.000 B/s
time_appconnect: 0.872926 s
time_connect: 0.282048 s
time_namelookup: 0.000524 s
time_pretransfer: 0.873099 s
time_redirect: 0.000000 s
time_starttransfer: 1.187467 s
time_total: 1452.904154 s

% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
2 66.6M 2 1809k 0 0 5760 0 3:22:15 0:05:21 3:16:54 0^C%

% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
52 66.6M 52 34.9M 0 0 16772 0 1:09:27 0:36:26 0:33:01 0^C%

% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
100 66.6M 100 66.6M 0 0 58181 0 0:20:01 0:20:01 --:--:-- 87975
http_code: 200
num_connects: 1
num_redirects: 0
remote_ip: 52.85.158.22
remote_port: 443
size_download: 69899433 B
speed_download: 58181.000 B/s
time_appconnect: 2.297713 s
time_connect: 1.904176 s
time_namelookup: 1.727602 s
time_pretransfer: 2.297974 s
time_redirect: 0.000000 s
time_starttransfer: 2.503263 s
time_total: 1201.408929 s
#+END_EXAMPLE


Well. As you see, the network in China is both slow and unstable.
G
G
Giovanni Biscuolo wrote on 19 Dec 2018 13:40
(name . Mark H Weaver)(address . mhw@netris.org)
87imzpd70s.fsf@roquette.mug.biscuolo.net
Hi Mark,

sorry for the late reply

Mark H Weaver <mhw@netris.org> writes:

Toggle quote (7 lines)
> Giovanni Biscuolo <g@xelera.eu> writes:
>> with a solid infrastructure of "scientifically" trustable build farms,
>> there are no reasons not to trust substitutes servers (this implies
>> working towards 100% reproducibility of GuixSD)
>
> What does "scientifically trustable" mean?

I'm still not able to elaborate on that (working on it, a sort of
self-research-hack project) but I'm referencing to this message related
to reduced bootstrap tarballs:


and the related reply by Jeremiah (unfortunately cannot find it in
archives, Message-ID: <877eh81tm4.fsf@ITSx01.pdp10.guru>)

in particular Jeremiah replied this:

Toggle snippet (9 lines)
> so, if I don't get it wrong, every skilled engineer will be able to
> build an "almost analogic" (zero bit of software preloaded) computing
> machine ad use stage0/mes [1] as the "metre" [2] to calibrate all other
> computing machines (thanks to reproducible builds)?

well, I haven't thought of it in those terms but yes I guess that is one
of the properties of the plan.

and

Toggle snippet (7 lines)
> so, having the scientific proof that binary conforms to source, there
> will be noo need to trust (the untrastable)

Well, that is what someone else could do with it but not a direct goal
of the work.

maybe a more correct definition of the above "scientific proof" should be
"mathematical proof"

I lack the theoretical basis to be more precise now, sorry :-S

a marketing-like campaign sould be "no more trusting trust"

best regards
Giovanni

--
Giovanni Biscuolo

Xelera IT Infrastructures
-----BEGIN PGP SIGNATURE-----
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=RMw8
-----END PGP SIGNATURE-----

M
M
Mark H Weaver wrote on 21 Dec 2018 01:23
Trustworthiness of build farms (was Re: CDN performance)
(name . Giovanni Biscuolo)(address . g@xelera.eu)
87fturwwup.fsf_-_@netris.org
Hi Giovanni,

Giovanni Biscuolo <g@xelera.eu> writes:

Toggle quote (40 lines)
> Mark H Weaver <mhw@netris.org> writes:
>
>> Giovanni Biscuolo <g@xelera.eu> writes:
>>> with a solid infrastructure of "scientifically" trustable build farms,
>>> there are no reasons not to trust substitutes servers (this implies
>>> working towards 100% reproducibility of GuixSD)
>>
>> What does "scientifically trustable" mean?
>
> I'm still not able to elaborate on that (working on it, a sort of
> self-research-hack project) but I'm referencing to this message related
> to reduced bootstrap tarballs:
>
> https://lists.gnu.org/archive/html/guix-devel/2018-11/msg00347.html
>
> and the related reply by Jeremiah (unfortunately cannot find it in
> archives, Message-ID: <877eh81tm4.fsf@ITSx01.pdp10.guru>)
>
> in particular Jeremiah replied this:
>
>> so, if I don't get it wrong, every skilled engineer will be able to
>> build an "almost analogic" (zero bit of software preloaded) computing
>> machine ad use stage0/mes [1] as the "metre" [2] to calibrate all other
>> computing machines (thanks to reproducible builds)?
>
> well, I haven't thought of it in those terms but yes I guess that is one
> of the properties of the plan.
>
>
> and
>
>> so, having the scientific proof that binary conforms to source, there
>> will be noo need to trust (the untrastable)
>
> Well, that is what someone else could do with it but not a direct goal
> of the work.
>
> maybe a more correct definition of the above "scientific proof" should be
> "mathematical proof"

I agree that a mathematical proof is what we should be aiming for, and
the only kind of proof that I could trust in, in this scenario.

However, it's important to note several caveats:

* A mathematical proof showing that the binary conforms to the source
requires a proof of correctness of the language implementation, which
in turn requires formal semantics for both the source language and the
underlying machine code. As far as I know, the current
bootstrappable.org effort does not include anything like this.
Existing projects that provide this include CakeML and Jitawa.

* One must assume that the hardware behaves according to its formal
specification. The veracity of this assumption is not something we
can currently verify, and even if we could, it would be invalidated if
someone gained physical access to the machine and modified it.

* The hardware initialization program (e.g. coreboot) and the kernel
used to perform the bootstrap must still be trusted, and unless I'm
mistaken, the bootstrappable.org effort does not currently address
these issues.

* The software running on the substitute servers could be compromised by
stealing SSH keys from someone with root access.

* If the private signing key of the substitute server can be stolen,
e.g. by gaining physical access to the machine for a short time, then
a man-in-the-middle can deliver to users compromised binaries that
appear to come from the substitute server itself.

* Not only the substitute server, but also all of its build slaves, must
be trusted as well.

So, while I certainly agree that it will be a *major* improvement to
avoid the need to blindly trust precompiled bootstrap binaries, we
should be clear that the current efforts fall far short of a proof, and
there still remain several valid reasons not to place one's trust in
substitute servers.

Does that make sense?

Regards,
Mark
C
C
Chris Marusich wrote on 21 Dec 2018 11:22
Re: CDN performance
(name . Meiyo Peng)(address . meiyo.peng@gmail.com)
87o99f9o2t.fsf@gmail.com
Hi Meiyo,

Thank you for sharing this information with us!

Can you also share what numbers you get when you run measure_get against
berlin.guixsd.org directly? Clearly, the connection from you to
CloudFront is not as performant as it is for others in other parts of
the world, but I wonder if it's still better than accessing berlin
directly. If you could run measure_get against berlin directly and
share the numbers, we can see if it represents any significant
improvement for you.

Meiyo Peng <meiyo.peng@gmail.com> writes:

Toggle quote (7 lines)
> I tested your script several times.
>
> 1. Tested today at home. China Unicom home broadband. 50Mb/s.
>
> The result is slow as usual. curl failed once.
> berlin-mirror.marusich.info is resolved to Seattle, WA, US.

Well, that's not great. Perhaps it's still better than it would be if
the DNS name resolved to a location in Europe, though.

Toggle quote (22 lines)
> #+BEGIN_EXAMPLE
> ? ~ measure_get https://berlin-mirror.marusich.info/nar/gzip/1bq783rbkzv9z9zdhivbvfzhsz2s5yac-linux-libre-4.19
> % Total % Received % Xferd Average Speed Time Time Time Current
> Dload Upload Total Spent Left Speed
> 55 66.6M 55 36.9M 0 0 17926 0 1:04:59 0:36:02 0:28:57 17733
> url_effective: https://berlin-mirror.marusich.info/nar/gzip/1bq783rbkzv9z9zdhivbvfzhsz2s5yac-linux-libre-4.19
> http_code: 200
> num_connects: 1
> num_redirects: 0
> remote_ip: 52.85.158.151
> remote_port: 443
> size_download: 38764357 B
> speed_download: 17926.000 B/s
> time_appconnect: 6.078850 s
> time_connect: 3.006821 s
> time_namelookup: 2.659785 s
> time_pretransfer: 6.079097 s
> time_redirect: 0.000000 s
> time_starttransfer: 9.626001 s
> time_total: 2162.379211 s
> curl: (92) HTTP/2 stream 0 was not closed cleanly: INTERNAL_ERROR (err 2)

I see, this is about 143 Kbps (not Mbps) of throughput, and 347 ms after
DNS name resolution.

Toggle quote (41 lines)
> % Total % Received % Xferd Average Speed Time Time Time Current
> Dload Upload Total Spent Left Speed
> 100 66.6M 100 66.6M 0 0 109k 0 0:10:25 0:10:25 --:--:-- 241k
> url_effective: https://berlin-mirror.marusich.info/nar/gzip/1bq783rbkzv9z9zdhivbvfzhsz2s5yac-linux-libre-4.19
> http_code: 200
> num_connects: 1
> num_redirects: 0
> remote_ip: 52.85.158.22
> remote_port: 443
> size_download: 69899433 B
> speed_download: 111816.000 B/s
> time_appconnect: 3.507528 s
> time_connect: 2.650373 s
> time_namelookup: 2.261801 s
> time_pretransfer: 3.507637 s
> time_redirect: 0.000000 s
> time_starttransfer: 5.995298 s
> time_total: 625.129571 s
>
> ? ~ measure_get https://berlin-mirror.marusich.info/nar/gzip/1bq783rbkzv9z9zdhivbvfzhsz2s5yac-linux-libre-4.19
> % Total % Received % Xferd Average Speed Time Time Time Current
> Dload Upload Total Spent Left Speed
> 100 66.6M 100 66.6M 0 0 109k 0 0:10:23 0:10:23 --:--:-- 141k
> url_effective: https://berlin-mirror.marusich.info/nar/gzip/1bq783rbkzv9z9zdhivbvfzhsz2s5yac-linux-libre-4.19
> http_code: 200
> num_connects: 1
> num_redirects: 0
> remote_ip: 52.85.158.22
> remote_port: 443
> size_download: 69899433 B
> speed_download: 112187.000 B/s
> time_appconnect: 2.280972 s
> time_connect: 1.407197 s
> time_namelookup: 1.056180 s
> time_pretransfer: 2.281234 s
> time_redirect: 0.000000 s
> time_starttransfer: 3.167703 s
> time_total: 623.061584 s
> #+END_EXAMPLE

897 Kbps, 351 ms after the name lookup.

Toggle quote (26 lines)
> 2. Tested 3 days ago at my office. China Telecom enterprise broadband. 50Mb/s.
>
> Unusually fast! berlin-mirror.marusich.info is resolved to Seattle, WA,
> US. I have no idea why it's so fast that day.
>
> #+BEGIN_EXAMPLE
> ? ~ measure_get https://berlin-mirror.marusich.info/nar/gzip/1bq783rbkzv9z9zdhivbvfzhsz2s5yac-linux-libre-4.19
> % Total % Received % Xferd Average Speed Time Time Time Current
> Dload Upload Total Spent Left Speed
> 100 66.6M 100 66.6M 0 0 1364k 0 0:00:50 0:00:50 --:--:-- 1352k
> url_effective: https://berlin-mirror.marusich.info/nar/gzip/1bq783rbkzv9z9zdhivbvfzhsz2s5yac-linux-libre-4.19
> http_code: 200
> num_connects: 1
> num_redirects: 0
> remote_ip: 13.35.20.109
> remote_port: 443
> size_download: 69899433 B
> speed_download: 1397429.000 B/s
> time_appconnect: 2.432387 s
> time_connect: 0.200842 s
> time_namelookup: 0.000446 s
> time_pretransfer: 2.432659 s
> time_redirect: 0.000000 s
> time_starttransfer: 2.673045 s
> time_total: 50.020945 s

11 Mbps, 200 ms after the name lookup.

Toggle quote (21 lines)
> % Total % Received % Xferd Average Speed Time Time Time Current
> Dload Upload Total Spent Left Speed
> 100 66.6M 100 66.6M 0 0 1592k 0 0:00:42 0:00:42 --:--:-- 2506k
> url_effective: https://berlin-mirror.marusich.info/nar/gzip/1bq783rbkzv9z9zdhivbvfzhsz2s5yac-linux-libre-4.19
> http_code: 200
> num_connects: 1
> num_redirects: 0
> remote_ip: 13.35.20.109
> remote_port: 443
> size_download: 69899433 B
> speed_download: 1630687.000 B/s
> time_appconnect: 0.653270 s
> time_connect: 0.209455 s
> time_namelookup: 0.001582 s
> time_pretransfer: 0.658399 s
> time_redirect: 0.000000 s
> time_starttransfer: 0.883126 s
> time_total: 42.865868 s
> #+END_EXAMPLE

13 Mbps, 208 ms after the name lookup.

Toggle quote (28 lines)
> 3. Tested today at my office. China Telecom enterprise broadband. 50Mb/s.
>
> Slow as usual. berlin-mirror.marusich.info is still resolved to Seattle,
> WA, US. I killed the program several times because it hung there with no
> data transfer for a few minutes. The TCP connection was probably closed
> by GFW. This is very common here.
>
> #+BEGIN_EXAMPLE
> ? ~ measure_get https://berlin-mirror.marusich.info/nar/gzip/1bq783rbkzv9z9zdhivbvfzhsz2s5yac-linux-libre-4.19
> % Total % Received % Xferd Average Speed Time Time Time Current
> Dload Upload Total Spent Left Speed
> 100 66.6M 100 66.6M 0 0 48110 0 0:24:12 0:24:12 --:--:-- 41808
> url_effective: https://berlin-mirror.marusich.info/nar/gzip/1bq783rbkzv9z9zdhivbvfzhsz2s5yac-linux-libre-4.19
> http_code: 200
> num_connects: 1
> num_redirects: 0
> remote_ip: 52.85.158.151
> remote_port: 443
> size_download: 69899433 B
> speed_download: 48110.000 B/s
> time_appconnect: 0.872926 s
> time_connect: 0.282048 s
> time_namelookup: 0.000524 s
> time_pretransfer: 0.873099 s
> time_redirect: 0.000000 s
> time_starttransfer: 1.187467 s
> time_total: 1452.904154 s

385 Kbps, 282 ms.

Toggle quote (30 lines)
> % Total % Received % Xferd Average Speed Time Time Time Current
> Dload Upload Total Spent Left Speed
> 2 66.6M 2 1809k 0 0 5760 0 3:22:15 0:05:21 3:16:54 0^C%
>
> ? ~ measure_get https://berlin-mirror.marusich.info/nar/gzip/1bq783rbkzv9z9zdhivbvfzhsz2s5yac-linux-libre-4.19
> % Total % Received % Xferd Average Speed Time Time Time Current
> Dload Upload Total Spent Left Speed
> 52 66.6M 52 34.9M 0 0 16772 0 1:09:27 0:36:26 0:33:01 0^C%
>
> ? ~ measure_get https://berlin-mirror.marusich.info/nar/gzip/1bq783rbkzv9z9zdhivbvfzhsz2s5yac-linux-libre-4.19
> % Total % Received % Xferd Average Speed Time Time Time Current
> Dload Upload Total Spent Left Speed
> 100 66.6M 100 66.6M 0 0 58181 0 0:20:01 0:20:01 --:--:-- 87975
> url_effective: https://berlin-mirror.marusich.info/nar/gzip/1bq783rbkzv9z9zdhivbvfzhsz2s5yac-linux-libre-4.19
> http_code: 200
> num_connects: 1
> num_redirects: 0
> remote_ip: 52.85.158.22
> remote_port: 443
> size_download: 69899433 B
> speed_download: 58181.000 B/s
> time_appconnect: 2.297713 s
> time_connect: 1.904176 s
> time_namelookup: 1.727602 s
> time_pretransfer: 2.297974 s
> time_redirect: 0.000000 s
> time_starttransfer: 2.503263 s
> time_total: 1201.408929 s

465 Kbps, 177 ms.

--
Chris
-----BEGIN PGP SIGNATURE-----
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=JveL
-----END PGP SIGNATURE-----

M
M
Meiyo Peng wrote on 21 Dec 2018 17:04
(name . Chris Marusich)(address . cmmarusich@gmail.com)
87y38isw77.fsf@gmail.com
Hi Chris,

Thank you for your patience!

Chris Marusich <cmmarusich@gmail.com> writes:

Toggle quote (8 lines)
> Can you also share what numbers you get when you run measure_get against
> berlin.guixsd.org directly? Clearly, the connection from you to
> CloudFront is not as performant as it is for others in other parts of
> the world, but I wonder if it's still better than accessing berlin
> directly. If you could run measure_get against berlin directly and
> share the numbers, we can see if it represents any significant
> improvement for you.

1. Tested today at home. China Unicom home broadband. 50Mb/s.

berlin.guixsd.org:

#+BEGIN_EXAMPLE
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
54 66.6M 54 36.3M 0 0 14981 0 1:17:45 0:42:25 0:35:20 0
http_code: 200
num_connects: 1
num_redirects: 0
remote_ip: 141.80.181.40
remote_port: 443
size_download: 38141765 B
speed_download: 14981.000 B/s
time_appconnect: 3.228601 s
time_connect: 2.213136 s
time_namelookup: 0.856194 s
time_pretransfer: 3.228820 s
time_redirect: 0.000000 s
time_starttransfer: 3.851583 s
time_total: 2545.889968 s
curl: (56) GnuTLS recv error (-54): Error in the pull function.

% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
100 66.6M 100 66.6M 0 0 20415 0 0:57:03 0:57:03 --:--:-- 25983
http_code: 200
num_connects: 1
num_redirects: 0
remote_ip: 141.80.181.40
remote_port: 443
size_download: 69899433 B
speed_download: 20415.000 B/s
time_appconnect: 2.005881 s
time_connect: 0.785257 s
time_namelookup: 0.000520 s
time_pretransfer: 2.006124 s
time_redirect: 0.000000 s
time_starttransfer: 3.031582 s
time_total: 3423.813489 s
#+END_EXAMPLE

berlin-mirror.marusich.info:

#+BEGIN_EXAMPLE
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
100 66.6M 100 66.6M 0 0 1470k 0 0:00:46 0:00:46 --:--:-- 2368k
http_code: 200
num_connects: 1
num_redirects: 0
remote_ip: 13.35.20.87
remote_port: 443
size_download: 69899433 B
speed_download: 1505934.000 B/s
time_appconnect: 3.343496 s
time_connect: 3.164926 s
time_namelookup: 3.060655 s
time_pretransfer: 3.343581 s
time_redirect: 0.000000 s
time_starttransfer: 5.766543 s
time_total: 46.416495 s

% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
100 66.6M 100 66.6M 0 0 3182k 0 0:00:21 0:00:21 --:--:-- 4612k
http_code: 200
num_connects: 1
num_redirects: 0
remote_ip: 13.35.20.87
remote_port: 443
size_download: 69899433 B
speed_download: 3259170.000 B/s
time_appconnect: 0.225982 s
time_connect: 0.070428 s
time_namelookup: 0.000483 s
time_pretransfer: 0.226055 s
time_redirect: 0.000000 s
time_starttransfer: 0.306621 s
time_total: 21.447966 s
#+END_EXAMPLE


2. Tested today at my office. China Telecom enterprise broadband. 50Mb/s.

berlin.guixsd.org:

#+BEGIN_EXAMPLE
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
100 66.6M 100 66.6M 0 0 3091k 0 0:00:22 0:00:22 --:--:-- 3649k
http_code: 200
num_connects: 1
num_redirects: 0
remote_ip: 141.80.181.40
remote_port: 443
size_download: 69899433 B
speed_download: 3166021.000 B/s
time_appconnect: 3.288213 s
time_connect: 2.733554 s
time_namelookup: 2.486754 s
time_pretransfer: 3.288320 s
time_redirect: 0.000000 s
time_starttransfer: 3.780341 s
time_total: 22.078489 s

% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
100 66.6M 100 66.6M 0 0 3499k 0 0:00:19 0:00:19 --:--:-- 4011k
http_code: 200
num_connects: 1
num_redirects: 0
remote_ip: 141.80.181.40
remote_port: 443
size_download: 69899433 B
speed_download: 3583667.000 B/s
time_appconnect: 0.761166 s
time_connect: 0.244415 s
time_namelookup: 0.000981 s
time_pretransfer: 0.761275 s
time_redirect: 0.000000 s
time_starttransfer: 1.247935 s
time_total: 19.505515 s
#+END_EXAMPLE

berlin-mirror.marusich.info:

#+BEGIN_EXAMPLE
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0
49 66.6M 49 32.8M 0 0 19012 0 1:01:16 0:30:13 0:31:03 29265
http_code: 200
num_connects: 1
num_redirects: 0
remote_ip: 52.85.158.22
remote_port: 443
size_download: 34488133 B
speed_download: 19012.000 B/s
time_appconnect: 2.958899 s
time_connect: 2.487483 s
time_namelookup: 2.271520 s
time_pretransfer: 2.959321 s
time_redirect: 0.000000 s
time_starttransfer: 5.447693 s
time_total: 1813.938029 s
curl: (92) HTTP/2 stream 0 was not closed cleanly: INTERNAL_ERROR (err 2)
#+END_EXAMPLE


Although both 13.35.20.0/24 and 52.85.158.0/24 IP ranges are located at Seattle,
the result shows that the connection to 13.35.20.0/24 is significantly faster.
This is pretty normal in China. It's definitely caused by the GFW. Giant
internet service providers (e.g. AWS) are the primary targets of the GFW.
M
M
Marius Bakke wrote on 21 Dec 2018 21:47
87zhsy8v5l.fsf@fastmail.com
Giovanni Biscuolo <g@xelera.eu> writes:

Toggle quote (18 lines)
> Mark H Weaver <mhw@netris.org> writes:
>
>> Giovanni Biscuolo <g@xelera.eu> writes:
>>> with a solid infrastructure of "scientifically" trustable build farms,
>>> there are no reasons not to trust substitutes servers (this implies
>>> working towards 100% reproducibility of GuixSD)
>>
>> What does "scientifically trustable" mean?
>
> I'm still not able to elaborate on that (working on it, a sort of
> self-research-hack project) but I'm referencing to this message related
> to reduced bootstrap tarballs:
>
> https://lists.gnu.org/archive/html/guix-devel/2018-11/msg00347.html
>
> and the related reply by Jeremiah (unfortunately cannot find it in
> archives, Message-ID: <877eh81tm4.fsf@ITSx01.pdp10.guru>)

-----BEGIN PGP SIGNATURE-----

iQEzBAEBCgAdFiEEu7At3yzq9qgNHeZDoqBt8qM6VPoFAlwdUUYACgkQoqBt8qM6
VPruRwgAhSOerLH74Awnfrly6rUDSJy3ujd7d2AY8etcSQwf83A5XJhVJ6w2TMRw
E1kE5aBeXDwL8tzfY0rvQC9BFMwouhpiR691ta8aGjXJ4XBN6VMQ2lstIR0i6jwb
hsGGApGw62ftaO54DAxTr0N35aedkvWEbGYydCAJ5Hpb5LjReeCL/AU/A8lbuVYA
n9KLfXuP8wv07Ss+xUdSudlDVsnnG+axXdBpP4mQl+556bsGldtz1QKGaN9l3wKR
pZlM01aDmda5ZLnh/up/5LoMFmWK7SAq+sqrmKf79rxUt45FvcB3p/Yz9XIaKF3s
JsMsd+SZs8m5xnM24A64jxvMUCAKLA==
=W5IM
-----END PGP SIGNATURE-----

R
R
Ricardo Wurmus wrote on 24 Dec 2018 15:47
(name . Giovanni Biscuolo)(address . g@xelera.eu)
87ftunq8wq.fsf@mdc-berlin.de
Hi Giovanni,

Toggle quote (12 lines)
> for this very reason IMHO we should work towards a network of **very
> trusted** build farms directly managed and controlled by the GuixSD
> project sysadmins; if build farms will be able to quickly provide
> substitutes, caching mirrors will be _much more_ effective than today
>
> ... and a network of "automated guix challenge" servers to spot
> not-reproducible software in GuixSD
>
> with a solid infrastructure of "scientifically" trustable build farms,
> there are no reasons not to trust substitutes servers (this implies
> working towards 100% reproducibility of GuixSD)

This sets the bar very high. I administer berlin.guix.info /
ci.guix.info (same server) and most of the associated build nodes, but I
don’t think people should trust these computers more than their own
computers or those managed by people they personally know.

The build servers do the same work that a user would do who builds
software locally without substitutes; the builders are supposed to
behave just like any other user. (And we can challenge their authority
with “guix challenge”.) I want us to keep in mind that build farms
don’t necessarily deserve any more trust than other computers you don’t
control. Substitute servers are a convenience.

To improve distribution of binaries we are committed to working on
different strategies at the same time:

- improve our processes so that non-critical package changes only hit
the master branch when we have already built the package and made it
available for distribution.

- improve availability of the single server berlin.guix.info by hooking
it up to a CDN via the name ci.guix.info (from which users can easily
opt out by changing their default substitute server to
berlin.guix.info).

- improve redundancy by setting up an off-site fail-over for the head of
the build farm at berlin.guix.info (such as bayfront).

- distribute build artefacts over IPFS without requiring a central
authority

All of these things can be done in parallel by different people. This
increases the project’s resilience and the options that users can choose
from.

--
Ricardo
P
P
Pierre Neidhardt wrote on 5 Mar 2019 17:19
unarchive 33600
(address . control@debbugs.gnu.org)
875zsxgupo.fsf@ambrevar.xyz
unarchive 33600
-----BEGIN PGP SIGNATURE-----

iQEzBAEBCAAdFiEEUPM+LlsMPZAEJKvom9z0l6S7zH8FAlx+oYMACgkQm9z0l6S7
zH+irQf/VdPjKDt+TYVxsxRjvDjdrUuPo0KKTPlpoRObBfxPJ+C/AQmn6AYcfoT1
vyey+lcSeFrnEfo1hIYlCT9T1TRIC+qIaK61BQVsMQV29o7wiJz2zSKw4bGDdGwk
2w7Dhcw8cM8V1mRQ5jppM4wtl0zq1AWwuQy2ehOOQ8H2Nt99zLYnbKC7GIqaSpZV
JeTmcc7gIoNEZeZOcsxhmsrICXgj7b+edyuZzTJbz21vV+XFFx/gIWlnYhpR8vZH
ws9m+YsRs1yxGNiNPdqH53ibKeY3l57hunzI6ICMS6+8qy+n/midjhN4Fql2CuKC
9kJBj1Z0wBVwLMgrAcyueVBekmyVZA==
=zIMj
-----END PGP SIGNATURE-----

?
Your comment

This issue is archived.

To comment on this conversation send an email to 33600@debbugs.gnu.org

To respond to this issue using the mumi CLI, first switch to it
mumi current 33600
Then, you may apply the latest patchset in this issue (with sign off)
mumi am -- -s
Or, compose a reply to this issue
mumi compose
Or, send patches to this issue
mumi send-email *.patch