Search Results: "mk"

29 September 2022

Antoine Beaupr : Detecting manual (and optimizing large) package installs in Puppet

Well this is a mouthful. I recently worked on a neat hack called puppet-package-check. It is designed to warn about manually installed packages, to make sure "everything is in Puppet". But it turns out it can (probably?) dramatically decrease the bootstrap time of Puppet bootstrap when it needs to install a large number of packages.

Detecting manual packages On a cleanly filed workstation, it looks like this:
root@emma:/home/anarcat/bin# ./puppet-package-check -v
listing puppet packages...
listing apt packages...
loading apt cache...
0 unmanaged packages found
A messy workstation will look like this:
root@curie:/home/anarcat/bin# ./puppet-package-check -v
listing puppet packages...
listing apt packages...
loading apt cache...
288 unmanaged packages found
apparmor-utils beignet-opencl-icd bridge-utils clustershell cups-pk-helper davfs2 dconf-cli dconf-editor dconf-gsettings-backend ddccontrol ddrescueview debmake debootstrap decopy dict-devil dict-freedict-eng-fra dict-freedict-eng-spa dict-freedict-fra-eng dict-freedict-spa-eng diffoscope dnsdiag dropbear-initramfs ebtables efibootmgr elpa-lua-mode entr eog evince figlet file file-roller fio flac flex font-manager fonts-cantarell fonts-inconsolata fonts-ipafont-gothic fonts-ipafont-mincho fonts-liberation fonts-monoid fonts-monoid-tight fonts-noto fonts-powerline fonts-symbola freeipmi freetype2-demos ftp fwupd-amd64-signed gallery-dl gcc-arm-linux-gnueabihf gcolor3 gcp gdisk gdm3 gdu gedit gedit-plugins gettext-base git-debrebase gnome-boxes gnote gnupg2 golang-any golang-docker-credential-helpers golang-golang-x-tools grub-efi-amd64-signed gsettings-desktop-schemas gsfonts gstreamer1.0-libav gstreamer1.0-plugins-base gstreamer1.0-plugins-good gstreamer1.0-plugins-ugly gstreamer1.0-pulseaudio gtypist gvfs-backends hackrf hashcat html2text httpie httping hugo humanfriendly iamerican-huge ibus ibus-gtk3 ibus-libpinyin ibus-pinyin im-config imediff img2pdf imv initramfs-tools input-utils installation-birthday internetarchive ipmitool iptables iptraf-ng jackd2 jupyter jupyter-nbextension-jupyter-js-widgets jupyter-qtconsole k3b kbtin kdialog keditbookmarks keepassxc kexec-tools keyboard-configuration kfind konsole krb5-locales kwin-x11 leiningen lightdm lintian linux-image-amd64 linux-perf lmodern lsb-base lvm2 lynx lz4json magic-wormhole mailscripts mailutils manuskript mat2 mate-notification-daemon mate-themes mime-support mktorrent mp3splt mpdris2 msitools mtp-tools mtree-netbsd mupdf nautilus nautilus-sendto ncal nd ndisc6 neomutt net-tools nethogs nghttp2-client nocache npm2deb ntfs-3g ntpdate nvme-cli nwipe obs-studio okular-extra-backends openstack-clients openstack-pkg-tools paprefs pass-extension-audit pcmanfm pdf-presenter-console pdf2svg percol pipenv playerctl plymouth plymouth-themes popularity-contest progress prometheus-node-exporter psensor pubpaste pulseaudio python3-ldap qjackctl qpdfview qrencode r-cran-ggplot2 r-cran-reshape2 rake restic rhash rpl rpm2cpio rs ruby ruby-dev ruby-feedparser ruby-magic ruby-mocha ruby-ronn rygel-playbin rygel-tracker s-tui sanoid saytime scrcpy scrcpy-server screenfetch scrot sdate sddm seahorse shim-signed sigil smartmontools smem smplayer sng sound-juicer sound-theme-freedesktop spectre-meltdown-checker sq ssh-audit sshuttle stress-ng strongswan strongswan-swanctl syncthing system-config-printer system-config-printer-common system-config-printer-udev systemd-bootchart systemd-container tardiff task-desktop task-english task-ssh-server tasksel tellico texinfo texlive-fonts-extra texlive-lang-cyrillic texlive-lang-french texlive-lang-german texlive-lang-italian texlive-xetex tftp-hpa thunar-archive-plugin tidy tikzit tint2 tintin++ tipa tpm2-tools traceroute tree trocla ucf udisks2 unifont unrar-free upower usbguard uuid-runtime vagrant-cachier vagrant-libvirt virt-manager vmtouch vorbis-tools w3m wamerican wamerican-huge wfrench whipper whohas wireshark xapian-tools xclip xdg-user-dirs-gtk xlax xmlto xsensors xserver-xorg xsltproc xxd xz-utils yubioath-desktop zathura zathura-pdf-poppler zenity zfs-dkms zfs-initramfs zfsutils-linux zip zlib1g zlib1g-dev
157 old: apparmor-utils clustershell davfs2 dconf-cli dconf-editor ddccontrol ddrescueview decopy dnsdiag ebtables efibootmgr elpa-lua-mode entr figlet file-roller fio flac flex font-manager freetype2-demos ftp gallery-dl gcc-arm-linux-gnueabihf gcolor3 gcp gdu gedit git-debrebase gnote golang-docker-credential-helpers golang-golang-x-tools gtypist hackrf hashcat html2text httpie httping hugo humanfriendly iamerican-huge ibus ibus-pinyin imediff input-utils internetarchive ipmitool iptraf-ng jackd2 jupyter-qtconsole k3b kbtin kdialog keditbookmarks keepassxc kexec-tools kfind konsole leiningen lightdm lynx lz4json magic-wormhole manuskript mat2 mate-notification-daemon mktorrent mp3splt msitools mtp-tools mtree-netbsd nautilus nautilus-sendto nd ndisc6 neomutt net-tools nethogs nghttp2-client nocache ntpdate nwipe obs-studio openstack-pkg-tools paprefs pass-extension-audit pcmanfm pdf-presenter-console pdf2svg percol pipenv playerctl qjackctl qpdfview qrencode r-cran-ggplot2 r-cran-reshape2 rake restic rhash rpl rpm2cpio rs ruby-feedparser ruby-magic ruby-mocha ruby-ronn s-tui saytime scrcpy screenfetch scrot sdate seahorse shim-signed sigil smem smplayer sng sound-juicer spectre-meltdown-checker sq ssh-audit sshuttle stress-ng system-config-printer system-config-printer-common tardiff tasksel tellico texlive-lang-cyrillic texlive-lang-french tftp-hpa tikzit tint2 tintin++ tpm2-tools traceroute tree unrar-free vagrant-cachier vagrant-libvirt vmtouch vorbis-tools w3m wamerican wamerican-huge wfrench whipper whohas xdg-user-dirs-gtk xlax xmlto xsensors xxd yubioath-desktop zenity zip
131 new: beignet-opencl-icd bridge-utils cups-pk-helper dconf-gsettings-backend debmake debootstrap dict-devil dict-freedict-eng-fra dict-freedict-eng-spa dict-freedict-fra-eng dict-freedict-spa-eng diffoscope dropbear-initramfs eog evince file fonts-cantarell fonts-inconsolata fonts-ipafont-gothic fonts-ipafont-mincho fonts-liberation fonts-monoid fonts-monoid-tight fonts-noto fonts-powerline fonts-symbola freeipmi fwupd-amd64-signed gdisk gdm3 gedit-plugins gettext-base gnome-boxes gnupg2 golang-any grub-efi-amd64-signed gsettings-desktop-schemas gsfonts gstreamer1.0-libav gstreamer1.0-plugins-base gstreamer1.0-plugins-good gstreamer1.0-plugins-ugly gstreamer1.0-pulseaudio gvfs-backends ibus-gtk3 ibus-libpinyin im-config img2pdf imv initramfs-tools installation-birthday iptables jupyter jupyter-nbextension-jupyter-js-widgets keyboard-configuration krb5-locales kwin-x11 lintian linux-image-amd64 linux-perf lmodern lsb-base lvm2 mailscripts mailutils mate-themes mime-support mpdris2 mupdf ncal npm2deb ntfs-3g nvme-cli okular-extra-backends openstack-clients plymouth plymouth-themes popularity-contest progress prometheus-node-exporter psensor pubpaste pulseaudio python3-ldap ruby ruby-dev rygel-playbin rygel-tracker sanoid scrcpy-server sddm smartmontools sound-theme-freedesktop strongswan strongswan-swanctl syncthing system-config-printer-udev systemd-bootchart systemd-container task-desktop task-english task-ssh-server texinfo texlive-fonts-extra texlive-lang-german texlive-lang-italian texlive-xetex thunar-archive-plugin tidy tipa trocla ucf udisks2 unifont upower usbguard uuid-runtime virt-manager wireshark xapian-tools xclip xserver-xorg xsltproc xz-utils zathura zathura-pdf-poppler zfs-dkms zfs-initramfs zfsutils-linux zlib1g zlib1g-dev
Yuck! That's a lot of shit to go through. Notice how the packages get sorted between "old" and "new" packages. This is because popcon is used as a tool to mark which packages are "old". If you have unmanaged packages, the "old" ones are likely things that you can uninstall, for example. If you don't have popcon installed, you'll also get this warning:
popcon stats not available: [Errno 2] No such file or directory: '/var/log/popularity-contest'
The error can otherwise be safely ignored, but you won't get "help" prioritizing the packages to add to your manifests. Note that the tool ignores packages that were "marked" (see apt-mark(8)) as automatically installed. This implies that you might have to do a little bit of cleanup the first time you run this, as Debian doesn't necessarily mark all of those packages correctly on first install. For example, here's how it looks like on a clean install, after Puppet ran:
root@angela:/home/anarcat# ./bin/puppet-package-check -v
listing puppet packages...
listing apt packages...
loading apt cache...
127 unmanaged packages found
ca-certificates console-setup cryptsetup-initramfs dbus file gcc-12-base gettext-base grub-common grub-efi-amd64 i3lock initramfs-tools iw keyboard-configuration krb5-locales laptop-detect libacl1 libapparmor1 libapt-pkg6.0 libargon2-1 libattr1 libaudit-common libaudit1 libblkid1 libbpf0 libbsd0 libbz2-1.0 libc6 libcap-ng0 libcap2 libcap2-bin libcom-err2 libcrypt1 libcryptsetup12 libdb5.3 libdebconfclient0 libdevmapper1.02.1 libedit2 libelf1 libext2fs2 libfdisk1 libffi8 libgcc-s1 libgcrypt20 libgmp10 libgnutls30 libgpg-error0 libgssapi-krb5-2 libhogweed6 libidn2-0 libip4tc2 libiw30 libjansson4 libjson-c5 libk5crypto3 libkeyutils1 libkmod2 libkrb5-3 libkrb5support0 liblocale-gettext-perl liblockfile-bin liblz4-1 liblzma5 libmd0 libmnl0 libmount1 libncurses6 libncursesw6 libnettle8 libnewt0.52 libnftables1 libnftnl11 libnl-3-200 libnl-genl-3-200 libnl-route-3-200 libnss-systemd libp11-kit0 libpam-systemd libpam0g libpcre2-8-0 libpcre3 libpcsclite1 libpopt0 libprocps8 libreadline8 libselinux1 libsemanage-common libsemanage2 libsepol2 libslang2 libsmartcols1 libss2 libssl1.1 libssl3 libstdc++6 libsystemd-shared libsystemd0 libtasn1-6 libtext-charwidth-perl libtext-iconv-perl libtext-wrapi18n-perl libtinfo6 libtirpc-common libtirpc3 libudev1 libunistring2 libuuid1 libxtables12 libxxhash0 libzstd1 linux-image-amd64 logsave lsb-base lvm2 media-types mlocate ncurses-term pass-extension-otp puppet python3-reportbug shim-signed tasksel ucf usr-is-merged util-linux-extra wpasupplicant xorg zlib1g
popcon stats not available: [Errno 2] No such file or directory: '/var/log/popularity-contest'
Normally, there should be unmanaged packages here. But because of the way Debian is installed, a lot of libraries and some core packages are marked as manually installed, and are of course not managed through Puppet. There are two solutions to this problem:
  • really manage everything in Puppet (argh)
  • mark packages as automatically installed
I typically chose the second path and mark a ton of stuff as automatic. Then either they will be auto-removed, or will stop being listed. In the above scenario, one could mark all libraries as automatically installed with:
apt-mark auto $(./bin/puppet-package-check   grep -o 'lib[^ ]*')
... but if you trust that most of that stuff is actually garbage that you don't really want installed anyways, you could just mark it all as automatically installed:
apt-mark auto $(./bin/puppet-package-check)
In my case, that ended up keeping basically all libraries (because of course they're installed for some reason) and auto-removing this:
dh-dkms discover-data dkms libdiscover2 libjsoncpp25 libssl1.1 linux-headers-amd64 mlocate pass-extension-otp pass-otp plocate x11-apps x11-session-utils xinit xorg
You'll notice xorg in there: yep, that's bad. Not what I wanted. But for some reason, on other workstations, I did not actually have xorg installed. Turns out having xserver-xorg is enough, and that one has dependencies. So now I guess I just learned to stop worrying and live without X(org).

Optimizing large package installs But that, of course, is not all. Why make things simple when you can have an unreadable title that is trying to be both syntactically correct and click-baity enough to flatter my vain ego? Right. One of the challenges in bootstrapping Puppet with large package lists is that it's slow. Puppet lists packages as individual resources and will basically run apt install $PKG on every package in the manifest, one at a time. While the overhead of apt is generally small, when you add things like apt-listbugs, apt-listchanges, needrestart, triggers and so on, it can take forever setting up a new host. So for initial installs, it can actually makes sense to skip the queue and just install everything in one big batch. And because the above tool inspects the packages installed by Puppet, you can run it against a catalog and have a full lists of all the packages Puppet would install, even before I even had Puppet running. So when reinstalling my laptop, I basically did this:
apt install puppet-agent/experimental
puppet agent --test --noop
apt install $(./puppet-package-check --debug \
    2>&1   grep ^puppet\ packages 
      sed 's/puppet packages://;s/ /\n/g'
      grep -v -e onionshare -e golint -e git-sizer -e github-backup -e hledger -e xsane -e audacity -e chirp -e elpa-flycheck -e elpa-lsp-ui -e yubikey-manager -e git-annex -e hopenpgp-tools -e puppet
) puppet-agent/experimental
That massive grep was because there are currently a lot of packages missing from bookworm. Those are all packages that I have in my catalog but that still haven't made it to bookworm. Sad, I know. I eventually worked around that by adding bullseye sources so that the Puppet manifest actually ran. The point here is that this improves the Puppet run time a lot. All packages get installed at once, and you get a nice progress bar. Then you actually run Puppet to deploy configurations and all the other goodies:
puppet agent --test
I wish I could tell you how much faster that ran. I don't know, and I will not go through a full reinstall just to please your curiosity. The only hard number I have is that it installed 444 packages (which exploded in 10,191 packages with dependencies) in a mere 10 minutes. That might also be with the packages already downloaded. In any case, I have that gut feeling it's faster, so you'll have to just trust my gut. It is, after all, much more important than you might think.

Similar work The blueprint system is something similar to this:
It figures out what you ve done manually, stores it locally in a Git repository, generates code that s able to recreate your efforts, and helps you deploy those changes to production
That tool has unfortunately been abandoned for a decade at this point. Also note that the AutoRemove::RecommendsImportant and AutoRemove::SuggestsImportant are relevant here. If it is set to true (the default), a package will not be removed if it is (respectively) a Recommends or Suggests of another package (as opposed to the normal Depends). In other words, if you want to also auto-remove packages that are only Suggests, you would, for example, add this to apt.conf:
AutoRemove::SuggestsImportant false;
Paul Wise has tried to make the Debian installer and debootstrap properly mark packages as automatically installed in the past, but his bug reports were rejected. The other suggestions in this section are also from Paul, thanks!

25 September 2022

Sergio Talens-Oliag: Kubernetes Static Content Server

This post describes how I ve put together a simple static content server for kubernetes clusters using a Pod with a persistent volume and multiple containers: an sftp server to manage contents, a web server to publish them with optional access control and another one to run scripts which need access to the volume filesystem. The sftp server runs using MySecureShell, the web server is nginx and the script runner uses the webhook tool to publish endpoints to call them (the calls will come from other Pods that run backend servers or are executed from Jobs or CronJobs).

HistoryThe system was developed because we had a NodeJS API with endpoints to upload files and store them on S3 compatible services that were later accessed via HTTPS, but the requirements changed and we needed to be able to publish folders instead of individual files using their original names and apply access restrictions using our API. Thinking about our requirements the use of a regular filesystem to keep the files and folders was a good option, as uploading and serving files is simple. For the upload I decided to use the sftp protocol, mainly because I already had an sftp container image based on mysecureshell prepared; once we settled on that we added sftp support to the API server and configured it to upload the files to our server instead of using S3 buckets. To publish the files we added a nginx container configured to work as a reverse proxy that uses the ngx_http_auth_request_module to validate access to the files (the sub request is configurable, in our deployment we have configured it to call our API to check if the user can access a given URL). Finally we added a third container when we needed to execute some tasks directly on the filesystem (using kubectl exec with the existing containers did not seem a good idea, as that is not supported by CronJobs objects, for example). The solution we found avoiding the NIH Syndrome (i.e. write our own tool) was to use the webhook tool to provide the endpoints to call the scripts; for now we have three:
  • one to get the disc usage of a PATH,
  • one to hardlink all the files that are identical on the filesystem,
  • one to copy files and folders from S3 buckets to our filesystem.

Container definitions

mysecureshellThe mysecureshell container can be used to provide an sftp service with multiple users (although the files are owned by the same UID and GID) using standalone containers (launched with docker or podman) or in an orchestration system like kubernetes, as we are going to do here. The image is generated using the following Dockerfile:
FROM alpine:$ALPINE_VERSION as builder
LABEL maintainer="Sergio Talens-Oliag <>"
RUN apk update &&\
 apk add --no-cache alpine-sdk git musl-dev &&\
 git clone &&\
 cd mysecureshell &&\
 ./configure --prefix=/usr --sysconfdir=/etc --mandir=/usr/share/man\
 --localstatedir=/var --with-shutfile=/var/lib/misc/sftp.shut --with-debug=2 &&\
 make all && make install &&\
 rm -rf /var/cache/apk/*
LABEL maintainer="Sergio Talens-Oliag <>"
COPY --from=builder /usr/bin/mysecureshell /usr/bin/mysecureshell
COPY --from=builder /usr/bin/sftp-* /usr/bin/
RUN apk update &&\
 apk add --no-cache openssh shadow pwgen &&\
 sed -i -e "s ^.*\(AuthorizedKeysFile\).*$ \1 /etc/ssh/auth_keys/%u "\
 /etc/ssh/sshd_config &&\
 mkdir /etc/ssh/auth_keys &&\
 cat /dev/null > /etc/motd &&\
 add-shell '/usr/bin/mysecureshell' &&\
 rm -rf /var/cache/apk/*
COPY bin/* /usr/local/bin/
COPY etc/sftp_config /etc/ssh/
VOLUME /sftp
CMD ["server"]
The /etc/sftp_config file is used to configure the mysecureshell server to have all the user homes under /sftp/data, only allow them to see the files under their home directories as if it were at the root of the server and close idle connections after 5m of inactivity:
# Default mysecureshell configuration
   # All users will have access their home directory under /sftp/data
   Home /sftp/data/$USER
   # Log to a file inside /sftp/logs/ (only works when the directory exists)
   LogFile /sftp/logs/mysecureshell.log
   # Force users to stay in their home directory
   StayAtHome true
   # Hide Home PATH, it will be shown as /
   VirtualChroot true
   # Hide real file/directory owner (just change displayed permissions)
   DirFakeUser true
   # Hide real file/directory group (just change displayed permissions)
   DirFakeGroup true
   # We do not want users to keep forever their idle connection
   IdleTimeOut 5m
# vim: ts=2:sw=2:et
The script is the one responsible to prepare the container for the users included on the /secrets/user_pass.txt file (creates the users with their HOME directories under /sftp/data and a /bin/false shell and creates the key files from /secrets/user_keys.txt if available). The script expects a couple of environment variables:
  • SFTP_UID: UID used to run the daemon and for all the files, it has to be different than 0 (all the files managed by this daemon are going to be owned by the same user and group, even if the remote users are different).
  • SFTP_GID: GID used to run the daemon and for all the files, it has to be different than 0.
And can use the SSH_PORT and SSH_PARAMS values if present. It also requires the following files (they can be mounted as secrets in kubernetes):
  • /secrets/host_keys.txt: Text file containing the ssh server keys in mime format; the file is processed using the reformime utility (the one included on busybox) and can be generated using the gen-host-keys script included on the container (it uses ssh-keygen and makemime).
  • /secrets/user_pass.txt: Text file containing lines of the form username:password_in_clear_text (only the users included on this file are available on the sftp server, in fact in our deployment we use only the scs user for everything).
And optionally can use another one:
  • /secrets/user_keys.txt: Text file that contains lines of the form username:public_ssh_ed25519_or_rsa_key; the public keys are installed on the server and can be used to log into the sftp server if the username exists on the user_pass.txt file.
The contents of the script are:
set -e
# ---------
# ---------
# Expects SSH_UID & SSH_GID on the environment and uses the value of the
# SSH_PORT & SSH_PARAMS variables if present
# Fixed values
HOST_KEY_TYPES="dsa ecdsa ed25519 rsa"
# ---------
# ---------
  # Check the ssh server keys ... we don't boot if we don't have them
  if [ ! -f "$HOST_KEYS" ]; then
    cat <<EOF
We need the host keys on the '$HOST_KEYS' file to proceed.
Call the 'gen-host-keys' script to create and export them on a mime file.
    exit 1
  # Check that we have users ... if we don't we can't continue
  if [ ! -f "$USER_PASS" ]; then
    cat <<EOF
We need at least the '$USER_PASS' file to provision users.
Call the 'gen-users-tar' script to create a tar file to create an archive that
contains public and private keys for users, a 'user_keys.txt' with the public
keys of the users and a 'user_pass.txt' file with random passwords for them 
(pass the list of usernames to it).
    exit 1
  # Check SFTP_UID
  if [ -z "$SFTP_UID" ]; then
    echo "The 'SFTP_UID' can't be empty, pass a 'GID'."
    exit 1
  if [ "$SFTP_UID" -eq "0" ]; then
    echo "The 'SFTP_UID' can't be 0, use a different 'UID'"
    exit 1
  # Check SFTP_GID
  if [ -z "$SFTP_GID" ]; then
    echo "The 'SFTP_GID' can't be empty, pass a 'GID'."
    exit 1
  if [ "$SFTP_GID" -eq "0" ]; then
    echo "The 'SFTP_GID' can't be 0, use a different 'GID'"
    exit 1
# Adjust ssh host keys
  tmpdir="$(mktemp -d)"
  cd "$tmpdir"
  reformime <"$HOST_KEYS"   ret="1"
  for kt in $HOST_KEY_TYPES; do
    key="ssh_host_$ kt _key"
    pub="ssh_host_$ kt"
    if [ ! -f "$key" ]; then
      echo "Missing '$key' file"
    if [ ! -f "$pub" ]; then
      echo "Missing '$pub' file"
    if [ "$ret" -ne "0" ]; then
    cat "$key" >"/etc/ssh/$key"
    chmod 0600 "/etc/ssh/$key"
    chown root:root "/etc/ssh/$key"
    cat "$pub" >"/etc/ssh/$pub"
    chmod 0600 "/etc/ssh/$pub"
    chown root:root "/etc/ssh/$pub"
  cd "$opwd"
  rm -rf "$tmpdir"
  return "$ret"
# Create users
  tmpdir="$(mktemp -d)"
  cd "$tmpdir"
  [ -d "$HOME_DIR" ]   mkdir "$HOME_DIR"
  # Make sure the data dir can be managed by the sftp user
  # Allow the user (and root) to create directories inside the $HOME_DIR, if
  # we don't allow it the directory creation fails on EFS (AWS)
  chmod 0755 "$HOME_DIR"
  # Create users
  echo "sftp:sftp:$SFTP_UID:$SFTP_GID:::/bin/false" >"newusers.txt"
  sed -n "/^[^#]/   s/:/ /p  " "$USER_PASS"   while read -r _u _p; do
    echo "$_u:$_p:$SFTP_UID:$SFTP_GID::$HOME_DIR/$_u:$USER_SHELL_CMD"
  done >>"newusers.txt"
  newusers --badnames newusers.txt
  # Disable write permission on the directory to forbid remote sftp users to
  # remove their own root dir (they have already done it); we adjust that
  # here to avoid issues with EFS (see before)
  chmod 0555 "$HOME_DIR"
  # Clean up the tmpdir
  cd "$opwd"
  rm -rf "$tmpdir"
  return "$ret"
# Adjust user keys
  if [ -f "$USER_KEYS" ]; then
    sed -n "/^[^#]/   s/:/ /p  " "$USER_KEYS"   while read -r _u _k; do
      echo "$_k" >>"$AUTH_KEYS_PATH/$_u"
# Main function
  echo "Running: /usr/sbin/sshd $SSH_PARAMS"
  # shellcheck disable=SC2086
  exec /usr/sbin/sshd -D $SSH_PARAMS
# ----
# ----
case "$1" in
"server") exec_sshd ;;
*) exec "$@" ;;
# vim: ts=2:sw=2:et
The container also includes a couple of auxiliary scripts, the first one can be used to generate the host_keys.txt file as follows:
$ docker run --rm stodh/mysecureshell gen-host-keys > host_keys.txt
Where the script is as simple as:
set -e
# Generate new host keys
ssh-keygen -A >/dev/null
# Replace hostname
sed -i -e 's/@.*$/@mysecureshell/' /etc/ssh/ssh_host_*
# Print in mime format (stdout)
makemime /etc/ssh/ssh_host_*
# vim: ts=2:sw=2:et
And there is another script to generate a .tar file that contains auth data for the list of usernames passed to it (the file contains a user_pass.txt file with random passwords for the users, public and private ssh keys for them and the user_keys.txt file that matches the generated keys). To generate a tar file for the user scs we can execute the following:
$ docker run --rm stodh/mysecureshell gen-users-tar scs > /tmp/scs-users.tar
To see the contents and the text inside the user_pass.txt file we can do:
$ tar tvf /tmp/scs-users.tar
-rw-r--r-- root/root        21 2022-09-11 15:55 user_pass.txt
-rw-r--r-- root/root       822 2022-09-11 15:55 user_keys.txt
-rw------- root/root       387 2022-09-11 15:55 id_ed25519-scs
-rw-r--r-- root/root        85 2022-09-11 15:55
-rw------- root/root      3357 2022-09-11 15:55 id_rsa-scs
-rw------- root/root      3243 2022-09-11 15:55 id_rsa-scs.pem
-rw-r--r-- root/root       729 2022-09-11 15:55
$ tar xfO /tmp/scs-users.tar user_pass.txt
The source of the script is:
set -e
# ---------
# ---------
# ---------
# ---------
# Generate user passwords and keys, return 1 if no username is received
if [ "$#" -eq "0" ]; then
  return 1
tmpdir="$(mktemp -d)"
cd "$tmpdir"
for u in "$@"; do
  ssh-keygen -q -a 100 -t ed25519 -f "id_ed25519-$u" -C "$u" -N ""
  ssh-keygen -q -a 100 -b 4096 -t rsa -f "id_rsa-$u" -C "$u" -N ""
  # Legacy RSA private key format
  cp -a "id_rsa-$u" "id_rsa-$u.pem"
  ssh-keygen -q -p -m pem -f "id_rsa-$u.pem" -N "" -P "" >/dev/null
  chmod 0600 "id_rsa-$u.pem"
  echo "$u:$(pwgen -s 16 1)" >>"$USER_PASS_FILE"
  echo "$u:$(cat "id_ed25519-$")" >>"$USER_KEYS_FILE"
  echo "$u:$(cat "id_rsa-$")" >>"$USER_KEYS_FILE"
tar cf - "$USER_PASS_FILE" "$USER_KEYS_FILE" id_* 2>/dev/null
cd "$opwd"
rm -rf "$tmpdir"
# vim: ts=2:sw=2:et

nginx-scsThe nginx-scs container is generated using the following Dockerfile:
LABEL maintainer="Sergio Talens-Oliag <>"
RUN rm -f /docker-entrypoint.d/*
COPY docker-entrypoint.d/* /docker-entrypoint.d/
Basically we are removing the existing docker-entrypoint.d scripts from the standard image and adding a new one that configures the web server as we want using a couple of environment variables:
  • AUTH_REQUEST_URI: URL to use for the auth_request, if the variable is not found on the environment auth_request is not used.
  • HTML_ROOT: Base directory of the web server, if not passed the default /usr/share/nginx/html is used.
Note that if we don t pass the variables everything works as if we were using the original nginx image. The contents of the configuration script are:
# Replace the default.conf nginx file by our own version.
set -e
if [ -z "$HTML_ROOT" ]; then
if [ "$AUTH_REQUEST_URI" ]; then
  cat >/etc/nginx/conf.d/default.conf <<EOF
  listen       80;
  server_name  localhost;
  location /  
    auth_request /.auth;
    root  $HTML_ROOT;
    index index.html index.htm;
  location /.auth  
    proxy_pass $AUTH_REQUEST_URI;
    proxy_pass_request_body off;
    proxy_set_header Content-Length "";
    proxy_set_header X-Original-URI \$request_uri;
  error_page   500 502 503 504  /50x.html;
  location = /50x.html  
    root /usr/share/nginx/html;
  cat >/etc/nginx/conf.d/default.conf <<EOF
  listen       80;
  server_name  localhost;
  location /  
    root  $HTML_ROOT;
    index index.html index.htm;
  error_page   500 502 503 504  /50x.html;
  location = /50x.html  
    root /usr/share/nginx/html;
# vim: ts=2:sw=2:et
As we will see later the idea is to use the /sftp/data or /sftp/data/scs folder as the root of the web published by this container and create an Ingress object to provide access to it outside of our kubernetes cluster.

webhook-scsThe webhook-scs container is generated using the following Dockerfile:
FROM golang:$GOLANG_VERSION AS builder
LABEL maintainer="Sergio Talens-Oliag <>"
WORKDIR /go/src/
RUN apk update &&\
 apk add --no-cache -t build-deps curl libc-dev gcc libgcc patch
RUN curl -L --silent -o webhook.tar.gz\$ WEBHOOK_VERSION .tar.gz &&\
 tar xzf webhook.tar.gz --strip 1 &&\
 curl -L --silent -o $ WEBHOOK_PR .patch\$ WEBHOOK_PR .patch &&\
 patch -p1 < $ WEBHOOK_PR .patch &&\
 go get -d && \
 go build -o /usr/local/bin/webhook
WORKDIR /src/s3fs-fuse
RUN apk update &&\
 apk add ca-certificates build-base alpine-sdk libcurl automake autoconf\
 libxml2-dev libressl-dev mailcap fuse-dev curl-dev
RUN curl -L --silent -o s3fs.tar.gz\$S3FS_VERSION.tar.gz &&\
 tar xzf s3fs.tar.gz --strip 1 &&\
 ./ &&\
 ./configure --prefix=/usr/local &&\
 make -j && \
 make install
LABEL maintainer="Sergio Talens-Oliag <>"
WORKDIR /webhook
RUN apk update &&\
 apk add --no-cache ca-certificates mailcap fuse libxml2 libcurl libgcc\
 libstdc++ rsync util-linux-misc &&\
 rm -rf /var/cache/apk/*
COPY --from=builder /usr/local/bin/webhook /usr/local/bin/webhook
COPY --from=builder /usr/local/bin/s3fs /usr/local/bin/s3fs
COPY hooks/* ./hooks/
CMD ["server"]
Again, we use a multi-stage build because in production we wanted to support a functionality that is not already on the official versions (streaming the command output as a response instead of waiting until the execution ends); this time we build the image applying the PATCH included on this pull request against a released version of the source instead of creating a fork. The script is used to generate the webhook configuration file for the existing hooks using environment variables (basically the WEBHOOK_WORKDIR and the *_TOKEN variables) and launch the webhook service:
set -e
# ---------
# ---------
WEBHOOK_BIN="$ WEBHOOK_BIN:-/webhook/hooks "
WEBHOOK_YML="$ WEBHOOK_YML:-/webhook/scs.yml "
# ---------
# ---------
  cat <<EOF
- id: du
  execute-command: '$WEBHOOK_BIN/'
  command-working-directory: '$WORKDIR'
  - name: 'Content-Type'
    value: 'application/json'
  http-methods: ['GET']
  include-command-output-in-response: true
  include-command-output-in-response-on-error: true
  - source: 'url'
    name: 'path'
  - source: 'string'
    envname: 'OUTPUT_FORMAT'
    name: 'json'
  cat <<EOF
- id: hardlink
  execute-command: '$WEBHOOK_BIN/'
  command-working-directory: '$WORKDIR'
  http-methods: ['GET']
  include-command-output-in-response: true
  include-command-output-in-response-on-error: true
  cat <<EOF
- id: s3sync
  execute-command: '$WEBHOOK_BIN/'
  command-working-directory: '$WORKDIR'
  http-methods: ['POST']
  include-command-output-in-response: true
  include-command-output-in-response-on-error: true
  - source: 'payload'
    envname: 'AWS_KEY'
    name: 'aws.key'
  - source: 'payload'
    envname: 'AWS_SECRET_KEY'
    name: 'aws.secret_key'
  - source: 'payload'
    envname: 'S3_BUCKET'
    name: 's3.bucket'
  - source: 'payload'
    envname: 'S3_REGION'
    name: 's3.region'
  - source: 'payload'
    envname: 'S3_PATH'
    name: 's3.path'
  - source: 'payload'
    envname: 'SCS_PATH'
    name: 'scs.path'
  stream-command-output: true
  if [ "$1" ]; then
    cat << EOF
      type: 'value'
      value: '$1'
        source: 'header'
        name: 'X-Webhook-Token'
  # Validate WORKDIR
  if [ -z "$WEBHOOK_WORKDIR" ]; then
    echo "Must define the WEBHOOK_WORKDIR variable!" >&2
    exit 1
  WORKDIR="$(realpath "$WEBHOOK_WORKDIR" 2>/dev/null)"   true
  if [ ! -d "$WORKDIR" ]; then
    echo "The WEBHOOK_WORKDIR '$WEBHOOK_WORKDIR' is not a directory!" >&2
    exit 1
  # Get TOKENS, if the DU_TOKEN or HARDLINK_TOKEN is defined that is used, if
  # not if the COMMON_TOKEN that is used and in other case no token is checked
  # (that is the default)
  # Create webhook configuration
    print_token_yml "$DU_TOKEN"
    echo ""
    print_token_yml "$HARDLINK_TOKEN"
    echo ""
    print_token_yml "$S3_TOKEN"
  # Run the webhook command
  # shellcheck disable=SC2086
  exec webhook -hooks "$WEBHOOK_YML" $WEBHOOK_OPTS
# ----
# ----
case "$1" in
"server") exec_webhook ;;
*) exec "$@" ;;
The script generates the configuration file for the webhook server calling functions that print a yaml section for each hook and optionally adds rules to validate access to them comparing the value of a X-Webhook-Token header against predefined values. The expected token values are taken from environment variables, we can define a token variable for each hook (DU_TOKEN, HARDLINK_TOKEN or S3_TOKEN) and a fallback value (COMMON_TOKEN); if no token variable is defined for a hook no check is done and everybody can call it. The Hook Definition documentation explains the options you can use for each hook, the ones we have right now do the following:
  • du: runs on the $WORKDIR directory, passes as first argument to the script the value of the path query parameter and sets the variable OUTPUT_FORMAT to the fixed value json (we use that to print the output of the script in JSON format instead of text).
  • hardlink: runs on the $WORKDIR directory and takes no parameters.
  • s3sync: runs on the $WORKDIR directory and sets a lot of environment variables from values read from the JSON encoded payload sent by the caller (all the values must be sent by the caller even if they are assigned an empty value, if they are missing the hook fails without calling the script); we also set the stream-command-output value to true to make the script show its output as it is working (we patched the webhook source to be able to use this option).

The du hook scriptThe du hook script code checks if the argument passed is a directory, computes its size using the du command and prints the results in text format or as a JSON dictionary:
set -e
# Script to print disk usage for a PATH inside the scs folder
# ---------
# ---------
  if [ "$OUTPUT_FORMAT" = "json" ]; then
    echo " \"error\":\"$*\" "
    echo "$*" >&2
  exit 1
  if [ "$OUTPUT_FORMAT" = "json" ]; then
    echo " \"error\":\"Pass arguments as '?path=XXX\" "
    echo "Usage: $(basename "$0") PATH" >&2
  exit 1
# ----
# ----
if [ "$#" -eq "0" ]   [ -z "$1" ]; then
if [ "$1" = "." ]; then
  DU_PATH="$(find . -name "$1" -mindepth 1 -maxdepth 1)"   true
if [ -z "$DU_PATH" ]   [ ! -d "$DU_PATH/." ]; then
  print_error "The provided PATH ('$1') is not a directory"
# Print disk usage in bytes for the given PATH
OUTPUT="$(du -b -s "$DU_PATH")"
if [ "$OUTPUT_FORMAT" = "json" ]; then
  # Format output as  "path":"PATH","bytes":"BYTES" 
  echo "$OUTPUT"  
    sed -e "s%^\(.*\)\t.*/\(.*\)$% \"path\":\"\2\",\"bytes\":\"\1\" %"  
    tr -d '\n'
  # Print du output as is
  echo "$OUTPUT"
# vim: ts=2:sw=2:et:ai:sts=2

The s3sync hook scriptThe s3sync hook script uses the s3fs tool to mount a bucket and synchronise data between a folder inside the bucket and a directory on the filesystem using rsync; all values needed to execute the task are taken from environment variables:
set -euo pipefail
set -o errexit
set -o errtrace
# Functions
  echo ""
  echo "Script exit code: $ret"
  exit "$ret"
# Check variables
if [ -z "$AWS_KEY" ]   [ -z "$AWS_SECRET_KEY" ]   [ -z "$S3_BUCKET" ]  
  [ -z "$S3_PATH" ]   [ -z "$SCS_PATH" ]; then
  [ "$AWS_KEY" ]   echo "Set the AWS_KEY environment variable"
  [ "$AWS_SECRET_KEY" ]   echo "Set the AWS_SECRET_KEY environment variable"
  [ "$S3_BUCKET" ]   echo "Set the S3_BUCKET environment variable"
  [ "$S3_PATH" ]   echo "Set the S3_PATH environment variable"
  [ "$SCS_PATH" ]   echo "Set the SCS_PATH environment variable"
  finish 1
if [ "$S3_REGION" ] && [ "$S3_REGION" != "us-east-1" ]; then
# Prepare working directory
WORK_DIR="$(mktemp -p "$HOME" -d)"
# Check the moutpoint
if [ ! -d "$MNT_POINT" ]; then
  mkdir -p "$MNT_POINT"
elif mountpoint "$MNT_POINT"; then
  echo "There is already something mounted on '$MNT_POINT', aborting!"
  finish 1
# Create password file
touch "$PASSWD_S3FS"
chmod 0400 "$PASSWD_S3FS"
# Mount s3 bucket as a filesystem
s3fs -o dbglevel=info,retries=5 -o "$EP_URL" -o "passwd_file=$PASSWD_S3FS" \
echo "Mounted bucket '$S3_BUCKET' on '$MNT_POINT'"
# Remove the password file, just in case
rm -f "$PASSWD_S3FS"
# Check source PATH
if [ ! -d "$SRC_PATH" ]; then
  echo "The S3_PATH '$S3_PATH' can't be found!"
# Compute SCS_UID & SCS_GID (by default based on the working directory owner)
SCS_UID="$ SCS_UID:=$(stat -c "%u" "." 2>/dev/null) "   true
SCS_GID="$ SCS_GID:=$(stat -c "%g" "." 2>/dev/null) "   true
# Check destination PATH
if [ "$ret" -eq "0" ] && [ -d "$DST_PATH" ]; then
  mkdir -p "$DST_PATH"   ret="$?"
# Copy using rsync
if [ "$ret" -eq "0" ]; then
  rsync -rlptv --chown="$SCS_UID:$SCS_GID" --delete --stats \
    "$SRC_PATH/" "$DST_PATH/"   ret="$?"
# Unmount the S3 bucket
umount -f "$MNT_POINT"
echo "Called umount for '$MNT_POINT'"
# Remove mount point dir
rmdir "$MNT_POINT"
# Remove WORK_DIR
rmdir "$WORK_DIR"
# We are done
finish "$ret"
# vim: ts=2:sw=2:et:ai:sts=2

Deployment objectsThe system is deployed as a StatefulSet with one replica. Our production deployment is done on AWS and to be able to scale we use EFS for our PersistenVolume; the idea is that the volume has no size limit, its AccessMode can be set to ReadWriteMany and we can mount it from multiple instances of the Pod without issues, even if they are in different availability zones. For development we use k3d and we are also able to scale the StatefulSet for testing because we use a ReadWriteOnce PVC, but it points to a hostPath that is backed up by a folder that is mounted on all the compute nodes, so in reality Pods in different k3d nodes use the same folder on the host.

secrets.yamlThe secrets file contains the files used by the mysecureshell container that can be generated using kubernetes pods as follows (we are only creating the scs user):
$ kubectl run "mysecureshell" --restart='Never' --quiet --rm --stdin \
  --image "stodh/mysecureshell:latest" -- gen-host-keys >"./host_keys.txt"
$ kubectl run "mysecureshell" --restart='Never' --quiet --rm --stdin \
  --image "stodh/mysecureshell:latest" -- gen-users-tar scs >"./users.tar"
Once we have the files we can generate the secrets.yaml file as follows:
$ tar xf ./users.tar user_keys.txt user_pass.txt
$ kubectl --dry-run=client -o yaml create secret generic "scs-secret" \
  --from-file="host_keys.txt=host_keys.txt" \
  --from-file="user_keys.txt=user_keys.txt" \
  --from-file="user_pass.txt=user_pass.txt" > ./secrets.yaml
The resulting secrets.yaml will look like the following file (the base64 would match the content of the files, of course):
apiVersion: v1
  host_keys.txt: TWlt...
  user_keys.txt: c2Nz...
  user_pass.txt: c2Nz...
kind: Secret
  creationTimestamp: null
  name: scs-secret

pvc.yamlThe persistent volume claim for a simple deployment (one with only one instance of the statefulSet) can be as simple as this:
apiVersion: v1
kind: PersistentVolumeClaim
  name: scs-pvc
  labels: scs
  - ReadWriteOnce
      storage: 8Gi
On this definition we don t set the storageClassName to use the default one.

Volumes in our development environment (k3d)In our development deployment we create the following PersistentVolume as required by the Local Persistence Volume Static Provisioner (note that the /volumes/scs-pv has to be created by hand, in our k3d system we mount the same host directory on the /volumes path of all the nodes and create the scs-pv directory by hand before deploying the persistent volume):
apiVersion: v1
kind: PersistentVolume
  name: scs-pv
  labels: scs
    storage: 8Gi
  volumeMode: Filesystem
  - ReadWriteOnce
  persistentVolumeReclaimPolicy: Delete
    name: scs-pvc
  storageClassName: local-storage
    path: /volumes/scs-pv
      - matchExpressions:
        - key:
          operator: In
          - k3s
And to make sure that everything works as expected we update the PVC definition to add the right storageClassName:
apiVersion: v1
kind: PersistentVolumeClaim
  name: scs-pvc
  labels: scs
  - ReadWriteOnce
      storage: 8Gi
  storageClassName: local-storage

Volumes in our production environment (aws)In the production deployment we don t create the PersistentVolume (we are using the aws-efs-csi-driver which supports Dynamic Provisioning) but we add the storageClassName (we set it to the one mapped to the EFS driver, i.e. efs-sc) and set ReadWriteMany as the accessMode:
apiVersion: v1
kind: PersistentVolumeClaim
  name: scs-pvc
  labels: scs
  - ReadWriteMany
      storage: 8Gi
  storageClassName: efs-sc

statefulset.yamlThe definition of the statefulSet is as follows:
apiVersion: apps/v1
kind: StatefulSet
  name: scs
  labels: scs
  serviceName: scs
  replicas: 1
      app: scs
        app: scs
      - name: nginx
        image: stodh/nginx-scs:latest
        - containerPort: 80
          name: http
        - name: AUTH_REQUEST_URI
          value: ""
        - name: HTML_ROOT
          value: /sftp/data
        - mountPath: /sftp
          name: scs-datadir
      - name: mysecureshell
        image: stodh/mysecureshell:latest
        - containerPort: 22
          name: ssh
            - IPC_OWNER
        - name: SFTP_UID
          value: '2020'
        - name: SFTP_GID
          value: '2020'
        - mountPath: /secrets
          name: scs-file-secrets
          readOnly: true
        - mountPath: /sftp
          name: scs-datadir
      - name: webhook
        image: stodh/webhook-scs:latest
          privileged: true
        - containerPort: 9000
          name: webhook-http
        - name: WEBHOOK_WORKDIR
          value: /sftp/data/scs
        - name: devfuse
          mountPath: /dev/fuse
        - mountPath: /sftp
          name: scs-datadir
      - name: devfuse
          path: /dev/fuse
      - name: scs-file-secrets
          secretName: scs-secrets
      - name: scs-datadir
          claimName: scs-pvc
Notes about the containers:
  • nginx: As this is an example the web server is not using an AUTH_REQUEST_URI and uses the /sftp/data directory as the root of the web (to get to the files uploaded for the scs user we will need to use /scs/ as a prefix on the URLs).
  • mysecureshell: We are adding the IPC_OWNER capability to the container to be able to use some of the sftp-* commands inside it, but they are not really needed, so adding the capability is optional.
  • webhook: We are launching this container in privileged mode to be able to use the s3fs-fuse, as it will not work otherwise for now (see this kubernetes issue); if the functionality is not needed the container can be executed with regular privileges; besides, as we are not enabling public access to this service we don t define *_TOKEN variables (if required the values should be read from a Secret object).
Notes about the volumes:
  • the devfuse volume is only needed if we plan to use the s3fs command on the webhook container, if not we can remove the volume definition and its mounts.

service.yamlTo be able to access the different services on the statefulset we publish the relevant ports using the following Service object:
apiVersion: v1
kind: Service
  name: scs-svc
  labels: scs
  - name: ssh
    port: 22
    protocol: TCP
    targetPort: 22
  - name: http
    port: 80
    protocol: TCP
    targetPort: 80
  - name: webhook-http
    port: 9000
    protocol: TCP
    targetPort: 9000
    app: scs

ingress.yamlTo download the scs files from the outside we can add an ingress object like the following (the definition is for testing using the localhost name):
kind: Ingress
  name: scs-ingress
  labels: scs
  ingressClassName: nginx
  - host: 'localhost'
      - path: /scs
        pathType: Prefix
            name: scs-svc
              number: 80

DeploymentTo deploy the statefulSet we create a namespace and apply the object definitions shown before:
$ kubectl create namespace scs-demo
namespace/scs-demo created
$ kubectl -n scs-demo apply -f secrets.yaml
secret/scs-secrets created
$ kubectl -n scs-demo apply -f pvc.yaml
persistentvolumeclaim/scs-pvc created
$ kubectl -n scs-demo apply -f statefulset.yaml
statefulset.apps/scs created
$ kubectl -n scs-demo apply -f service.yaml
service/scs-svc created
$ kubectl -n scs-demo apply -f ingress.yaml created
Once the objects are deployed we can check that all is working using kubectl:
$ kubectl  -n scs-demo get all,secrets,ingress
pod/scs-0   3/3     Running   0          24s
NAME            TYPE       CLUSTER-IP  EXTERNAL-IP  PORT(S)                  AGE
service/scs-svc ClusterIP  <none>       22/TCP,80/TCP,9000/TCP   21s

NAME                   READY   AGE
statefulset.apps/scs   1/1     24s
NAME                         TYPE                                  DATA   AGE
secret/default-token-mwcd7   3      53s
secret/scs-secrets           Opaque                                3      39s
NAME                                   CLASS  HOSTS      ADDRESS     PORTS   AGE  nginx  localhost  80      17s
At this point we are ready to use the system.

Usage examples

File uploadsAs previously mentioned in our system the idea is to use the sftp server from other Pods, but to test the system we are going to do a kubectl port-forward and connect to the server using our host client and the password we have generated (it is on the user_pass.txt file, inside the users.tar archive):
$ kubectl -n scs-demo port-forward service/scs-svc 2020:22 &
Forwarding from -> 22
Forwarding from [::1]:2020 -> 22
$ PF_PID=$!
$ sftp -P 2020 scs@                                                 1
Handling connection for 2020
The authenticity of host '[]:2020 ([]:2020)' can't be \
ED25519 key fingerprint is SHA256:eHNwCnyLcSSuVXXiLKeGraw0FT/4Bb/yjfqTstt+088.
This key is not known by any other names
Are you sure you want to continue connecting (yes/no/[fingerprint])? yes
Warning: Permanently added '[]:2020' (ED25519) to the list of known \
scs@'s password: **********
Connected to
sftp> ls -la
drwxr-xr-x    2 sftp     sftp         4096 Sep 25 14:47 .
dr-xr-xr-x    3 sftp     sftp         4096 Sep 25 14:36 ..
sftp> !date -R > /tmp/date.txt                                               2
sftp> put /tmp/date.txt .
Uploading /tmp/date.txt to /date.txt
date.txt                                      100%   32    27.8KB/s   00:00
sftp> ls -l
-rw-r--r--    1 sftp     sftp           32 Sep 25 15:21 date.txt
sftp> ln date.txt date.txt.1                                                 3
sftp> ls -l
-rw-r--r--    2 sftp     sftp           32 Sep 25 15:21 date.txt
-rw-r--r--    2 sftp     sftp           32 Sep 25 15:21 date.txt.1
sftp> put /tmp/date.txt date.txt.2                                           4
Uploading /tmp/date.txt to /date.txt.2
date.txt                                      100%   32    27.8KB/s   00:00
sftp> ls -l                                                                  5
-rw-r--r--    2 sftp     sftp           32 Sep 25 15:21 date.txt
-rw-r--r--    2 sftp     sftp           32 Sep 25 15:21 date.txt.1
-rw-r--r--    1 sftp     sftp           32 Sep 25 15:21 date.txt.2
sftp> exit
$ kill "$PF_PID"
[1]  + terminated  kubectl -n scs-demo port-forward service/scs-svc 2020:22
  1. We connect to the sftp service on the forwarded port with the scs user.
  2. We put a file we have created on the host on the directory.
  3. We do a hard link of the uploaded file.
  4. We put a second copy of the file we created locally.
  5. On the file list we can see that the two first files have two hardlinks

File retrievalsIf our ingress is configured right we can download the date.txt file from the URL http://localhost/scs/date.txt:
$ curl -s http://localhost/scs/date.txt
Sun, 25 Sep 2022 17:21:51 +0200

Use of the webhook containerTo finish this post we are going to show how we can call the hooks directly, from a CronJob and from a Job.

Direct script call (du)In our deployment the direct calls are done from other Pods, to simulate it we are going to do a port-forward and call the script with an existing PATH (the root directory) and a bad one:
$ kubectl -n scs-demo port-forward service/scs-svc 9000:9000 >/dev/null &
$ PF_PID=$!
$ JSON="$(curl -s "http://localhost:9000/hooks/du?path=.")"
$ echo $JSON
$ JSON="$(curl -s "http://localhost:9000/hooks/du?path=foo")"
$ echo $JSON
 "error":"The provided PATH ('foo') is not a directory" 
$ kill $PF_PID
As we only have files on the base directory we print the disk usage of the . PATH and the output is in json format because we export OUTPUT_FORMAT with the value json on the webhook configuration.

Jobs (s3sync)The following job can be used to synchronise the contents of a directory in a S3 bucket with the SCS Filesystem:
apiVersion: batch/v1
kind: Job
  name: s3sync
    cronjob: 's3sync'
        cronjob: 's3sync'
      - name: s3sync-job
        image: alpine:latest
        - "wget"
        - "-q"
        - "--header"
        - "Content-Type: application/json"
        - "--post-file"
        - "/secrets/s3sync.json"
        - "-O-"
        - "http://scs-svc:9000/hooks/s3sync"
        - mountPath: /secrets
          name: job-secrets
          readOnly: true
      restartPolicy: Never
      - name: job-secrets
          secretName: webhook-job-secrets
The file with parameters for the script must be something like this:
    "key": "********************",
    "secret_key": "****************************************"
    "region": "eu-north-1",
    "bucket": "blogops-test",
    "path": "test"
    "path": "test"
Once we have both files we can run the Job as follows:
$ kubectl -n scs-demo create secret generic webhook-job-secrets \            1
secret/webhook-job-secrets created
$ kubectl -n scs-demo apply -f webhook-job.yaml                              2
job.batch/s3sync created
$ kubectl -n scs-demo get pods -l "cronjob=s3sync"                           3
s3sync-zx2cj   0/1     Completed   0          12s
$ kubectl -n scs-demo logs s3sync-zx2cj                                      4
Mounted bucket 's3fs-test' on '/root/tmp.jiOjaF/s3data'
sending incremental file list
created directory ./test
Number of files: 2 (reg: 1, dir: 1)
Number of created files: 2 (reg: 1, dir: 1)
Number of deleted files: 0
Number of regular files transferred: 1
Total file size: 15,075 bytes
Total transferred file size: 15,075 bytes
Literal data: 15,075 bytes
Matched data: 0 bytes
File list size: 0
File list generation time: 0.147 seconds
File list transfer time: 0.000 seconds
Total bytes sent: 15,183
Total bytes received: 74
sent 15,183 bytes  received 74 bytes  30,514.00 bytes/sec
total size is 15,075  speedup is 0.99
Called umount for '/root/tmp.jiOjaF/s3data'
Script exit code: 0
$ kubectl -n scs-demo delete -f webhook-job.yaml                             5
job.batch "s3sync" deleted
$ kubectl -n scs-demo delete secrets webhook-job-secrets                     6
secret "webhook-job-secrets" deleted
  1. Here we create the webhook-job-secrets secret that contains the s3sync.json file.
  2. This command runs the job.
  3. Checking the label cronjob=s3sync we get the Pods executed by the job.
  4. Here we print the logs of the completed job.
  5. Once we are finished we remove the Job.
  6. And also the secret.

Final remarksThis post has been longer than I expected, but I believe it can be useful for someone; in any case, next time I ll try to explain something shorter or will split it into multiple entries.

20 September 2022

Simon Josefsson: Privilege separation of GSS-API credentials for Apache

To protect web resources with Kerberos you may use Apache HTTPD with mod_auth_gssapi however, all web scripts (e.g., PHP) run under Apache will have access to the Kerberos long-term symmetric secret credential (keytab). If someone can get it, they can impersonate your server, which is bad. The gssproxy project makes it possible to introduce privilege separation to reduce the attack surface. There is a tutorial for RPM-based distributions (Fedora, RHEL, AlmaLinux, etc), but I wanted to get this to work on a DPKG-based distribution (Debian, Ubuntu, Trisquel, PureOS, etc) and found it worthwhile to document the process. I m using Ubuntu 22.04 below, but have tested it on Debian 11 as well. I have adopted the gssproxy package in Debian, and testing this setup is part of the scripted autopkgtest/debci regression testing. First install the required packages:
root@foo:~# apt-get update
root@foo:~# apt-get install -y apache2 libapache2-mod-auth-gssapi gssproxy curl
This should give you a working and running web server. Verify it is operational under the proper hostname, I ll use in this writeup.
root@foo:~# curl --head
HTTP/1.1 200 OK
The next step is to create a keytab containing the Kerberos V5 secrets for your host, the exact steps depends on your environment (usually kadmin ktadd or ipa-getkeytab), but use the string HTTP/ and then confirm using something like the following.
root@foo:~# ls -la /etc/gssproxy/httpd.keytab
-rw------- 1 root root 176 Sep 18 06:44 /etc/gssproxy/httpd.keytab
root@foo:~# klist -k /etc/gssproxy/httpd.keytab -e
Keytab name: FILE:/etc/gssproxy/httpd.keytab
KVNO Principal
---- --------------------------------------------------------------------------
   2 HTTP/ (aes256-cts-hmac-sha1-96) 
   2 HTTP/ (aes128-cts-hmac-sha1-96) 
The file should be owned by root and not be in the default /etc/krb5.keytab location, so Apache s libapache2-mod-auth-gssapi will have to use gssproxy to use it.

Then configure gssproxy to find the credential and use it with Apache.
root@foo:~# cat<<EOF > /etc/gssproxy/80-httpd.conf
mechs = krb5
cred_store = keytab:/etc/gssproxy/httpd.keytab
cred_store = ccache:/var/lib/gssproxy/clients/krb5cc_%U
euid = www-data
process = /usr/sbin/apache2
For debugging, it may be useful to enable more gssproxy logging:
root@foo:~# cat<<EOF > /etc/gssproxy/gssproxy.conf
debug_level = 1
Restart gssproxy so it finds the new configuration, and monitor syslog as follows:
root@foo:~# tail -F /var/log/syslog &
root@foo:~# systemctl restart gssproxy
You should see something like this in the log file:
Sep 18 07:03:15 foo gssproxy[4076]: [2022/09/18 05:03:15]: Exiting after receiving a signal
Sep 18 07:03:15 foo systemd[1]: Stopping GSSAPI Proxy Daemon
Sep 18 07:03:15 foo systemd[1]: gssproxy.service: Deactivated successfully.
Sep 18 07:03:15 foo systemd[1]: Stopped GSSAPI Proxy Daemon.
Sep 18 07:03:15 foo gssproxy[4092]: [2022/09/18 05:03:15]: Debug Enabled (level: 1)
Sep 18 07:03:15 foo systemd[1]: Starting GSSAPI Proxy Daemon
Sep 18 07:03:15 foo gssproxy[4093]: [2022/09/18 05:03:15]: Kernel doesn't support GSS-Proxy (can't open /proc/net/rpc/use-gss-proxy: 2 (No such file or directory))
Sep 18 07:03:15 foo gssproxy[4093]: [2022/09/18 05:03:15]: Problem with kernel communication! NFS server will not work
Sep 18 07:03:15 foo systemd[1]: Started GSSAPI Proxy Daemon.
Sep 18 07:03:15 foo gssproxy[4093]: [2022/09/18 05:03:15]: Initialization complete.
The NFS-related errors is due to a default gssproxy configuration file, it is harmless and if you don t use NFS with GSS-API you can silence it like this:
root@foo:~# rm /etc/gssproxy/24-nfs-server.conf
root@foo:~# systemctl try-reload-or-restart gssproxy
The log should now indicate that it loaded the keytab:
Sep 18 07:18:59 foo systemd[1]: Reloading GSSAPI Proxy Daemon 
Sep 18 07:18:59 foo gssproxy[4182]: [2022/09/18 05:18:59]: Received SIGHUP; re-reading config.
Sep 18 07:18:59 foo gssproxy[4182]: [2022/09/18 05:18:59]: Service: HTTP, Keytab: /etc/gssproxy/httpd.keytab, Enctype: 18
Sep 18 07:18:59 foo gssproxy[4182]: [2022/09/18 05:18:59]: New config loaded successfully.
Sep 18 07:18:59 foo systemd[1]: Reloaded GSSAPI Proxy Daemon.
To instruct Apache or actually, the MIT Kerberos V5 GSS-API library used by mod_auth_gssap loaded by Apache to use gssproxy instead of using /etc/krb5.keytab as usual, Apache needs to be started in an environment that has GSS_USE_PROXY=1 set. The background is covered by the gssproxy-mech(8) man page and explained by the gssproxy README.

When systemd is used the following can be used to set the environment variable, note the final command to reload systemd.
root@foo:~# mkdir -p /etc/systemd/system/apache2.service.d
root@foo:~# cat<<EOF > /etc/systemd/system/apache2.service.d/gssproxy.conf
root@foo:~# systemctl daemon-reload
The next step is to configure a GSS-API protected Apache resource:
root@foo:~# cat<<EOF > /etc/apache2/conf-available/private.conf
<Location /private>
  AuthType GSSAPI
  AuthName "GSSAPI Login"
  Require valid-user
Enable the configuration and restart Apache the suggested use of reload is not sufficient, because then it won t be restarted with the newly introduced GSS_USE_PROXY variable. This just applies to the first time, after the first restart you may use reload again.
root@foo:~# a2enconf private
Enabling conf private.
To activate the new configuration, you need to run:
systemctl reload apache2
root@foo:~# systemctl restart apache2
When you have debug messages enabled, the log may look like this:
Sep 18 07:32:23 foo systemd[1]: Stopping The Apache HTTP Server 
Sep 18 07:32:23 foo gssproxy[4182]: [2022/09/18 05:32:23]: Client [2022/09/18 05:32:23]: (/usr/sbin/apache2) [2022/09/18 05:32:23]: connected (fd = 10)[2022/09/18 05:32:23]: (pid = 4651) (uid = 0) (gid = 0)[2022/09/18 05:32:23]:
Sep 18 07:32:23 foo gssproxy[4182]: message repeated 4 times: [ [2022/09/18 05:32:23]: Client [2022/09/18 05:32:23]: (/usr/sbin/apache2) [2022/09/18 05:32:23]: connected (fd = 10)[2022/09/18 05:32:23]: (pid = 4651) (uid = 0) (gid = 0)[2022/09/18 05:32:23]:]
Sep 18 07:32:23 foo systemd[1]: apache2.service: Deactivated successfully.
Sep 18 07:32:23 foo systemd[1]: Stopped The Apache HTTP Server.
Sep 18 07:32:23 foo systemd[1]: Starting The Apache HTTP Server
Sep 18 07:32:23 foo gssproxy[4182]: [2022/09/18 05:32:23]: Client [2022/09/18 05:32:23]: (/usr/sbin/apache2) [2022/09/18 05:32:23]: connected (fd = 10)[2022/09/18 05:32:23]: (pid = 4657) (uid = 0) (gid = 0)[2022/09/18 05:32:23]:
root@foo:~# Sep 18 07:32:23 foo gssproxy[4182]: message repeated 8 times: [ [2022/09/18 05:32:23]: Client [2022/09/18 05:32:23]: (/usr/sbin/apache2) [2022/09/18 05:32:23]: connected (fd = 10)[2022/09/18 05:32:23]: (pid = 4657) (uid = 0) (gid = 0)[2022/09/18 05:32:23]:]
Sep 18 07:32:23 foo systemd[1]: Started The Apache HTTP Server.
Finally, set up a dummy test page on the server:
root@foo:~# echo OK > /var/www/html/private
To verify that the server is working properly you may acquire tickets locally and then use curl to retrieve the GSS-API protected resource. The "--negotiate" enables SPNEGO and "--user :" asks curl to use username from the environment.
root@foo:~# klist
Ticket cache: FILE:/tmp/krb5cc_0
Default principal: jas@GSSPROXY.EXAMPLE.ORG
Valid starting Expires Service principal
09/18/22 07:40:37 09/19/22 07:40:37 krbtgt/GSSPROXY.EXAMPLE.ORG@GSSPROXY.EXAMPLE.ORG
root@foo:~# curl --negotiate --user :
The log should contain something like this:
Sep 18 07:56:00 foo gssproxy[4872]: [2022/09/18 05:56:00]: Client [2022/09/18 05:56:00]: (/usr/sbin/apache2) [2022/09/18 05:56:00]: connected (fd = 10)[2022/09/18 05:56:00]: (pid = 5042) (uid = 33) (gid = 33)[2022/09/18 05:56:00]:
Sep 18 07:56:00 foo gssproxy[4872]: [CID 10][2022/09/18 05:56:00]: gp_rpc_execute: executing 6 (GSSX_ACQUIRE_CRED) for service "HTTP", euid: 33,socket: (null)
Sep 18 07:56:00 foo gssproxy[4872]: [CID 10][2022/09/18 05:56:00]: gp_rpc_execute: executing 6 (GSSX_ACQUIRE_CRED) for service "HTTP", euid: 33,socket: (null)
Sep 18 07:56:00 foo gssproxy[4872]: [CID 10][2022/09/18 05:56:00]: gp_rpc_execute: executing 1 (GSSX_INDICATE_MECHS) for service "HTTP", euid: 33,socket: (null)
Sep 18 07:56:00 foo gssproxy[4872]: [CID 10][2022/09/18 05:56:00]: gp_rpc_execute: executing 6 (GSSX_ACQUIRE_CRED) for service "HTTP", euid: 33,socket: (null)
Sep 18 07:56:00 foo gssproxy[4872]: [CID 10][2022/09/18 05:56:00]: gp_rpc_execute: executing 9 (GSSX_ACCEPT_SEC_CONTEXT) for service "HTTP", euid: 33,socket: (null)
The Apache log will look like this, notice the authenticated username shown. - jas@GSSPROXY.EXAMPLE.ORG [18/Sep/2022:07:56:00 +0200] "GET /private HTTP/1.1" 200 481 "-" "curl/7.81.0"
Congratulations, and happy hacking!

15 September 2022

Joachim Breitner: rec-def: Dominators case study

More ICFP-inspired experiments using the rec-def library: In Norman Ramsey s very nice talk about his Functional Pearl Beyond Relooper: Recursive Translation of Unstructured Control Flow to Structured Control Flow , he had the following slide showing the equation for the dominators of a node in a graph:
Norman Ramsey shows a formula Norman Ramsey shows a formula
He said it s ICFP and I wanted to say the dominance relation has a beautiful set of equations you can read all these algorithms how to compute this, but the concept is simple . This made me wonder: If the concept is simple and this formula is beautiful shouldn t this be sufficient for the Haskell programmer to obtain the dominator relation, without reading all those algorithms? Before we start, we have to clarify the formula a bit: If a node is an entry node (no predecessors) then the big intersection is over the empty set, and that is not a well-defined concept. For these nodes, we need that big intersection to return the empty set, as entry nodes are not dominated by any other node. (Let s assume that the entry nodes are exactly those with no predecessors.) Let s try, first using plain Haskell data structures. We begin by implementing this big intersection operator on Data.Set, and also a function to find the predecessors of a node in a graph: Now we can write down the formula that Norman gave, quite elegantly: Does this work? It seems it does: But not surprising if you have read my previous blog posts it falls over once we have recursion: So let us reimplement it with Data.Recursive.Set. The hope is that we can simply replace the operations, and that now it can suddenly handle cyclic graphs as well. Let s see: It does! Well, it does return a result but it looks strange. Clearly node 3 and 4 are also dominated by 1, but the result does not reflect that. But the result is a solution to Norman s equation. Was the equation wrong? No, but we failed to notice that the desired solution is the largest, not the smallest. And Data.Recursive.Set calculates, as documented, the least fixed point. What now? Until the library has code for RDualSet a, we can work around this by using the dual formula to calculate the non-dominators. To do this, we
  • use union instead of intersection
  • delete instead of insert,
  • S.empty, use the set of all nodes (which requires some extra plumbing)
  • subtract the result from the set of all nodes to get the dominators
and thus the code turns into:
And with this, now we do get the correct result:
ghci> domintors3 [(1,2),(1,3),(2,4),(3,4),(4,3)]
fromList [(1,[1]),(2,[1,2]),(3,[1,3]),(4,[1,4])]
We worked a little bit on how to express the beautiful formula to Haskell, but at no point did we have to think about how to solve it. To me, this is the essence of declarative programming.

13 September 2022

Alberto Garc a: Adding software to the Steam Deck with systemd-sysext

Yakuake on SteamOS Introduction: an immutable OS The Steam Deck runs SteamOS, a single-user operating system based on Arch Linux. Although derived from a standard package-based distro, the OS in the Steam Deck is immutable and system updates replace the contents of the root filesystem atomically instead of using the package manager. An immutable OS makes the system more stable and its updates less error-prone, but users cannot install additional packages to add more software. This is not a problem for most users since they are only going to run Steam and its games (which are stored in the home partition). Nevertheless, the OS also has a desktop mode which provides a standard Linux desktop experience, and here it makes sense to be able to install more software. How to do that though? It is possible for the user to become root, make the root filesytem read-write and install additional software there, but any changes will be gone after the next OS update. Modifying the rootfs can also be dangerous if the user is not careful. Ways to add additional software The simplest and safest way to install additional software is with Flatpak, and that s the method recommended in the Steam Deck Desktop FAQ. Flatpak is already installed and integrated in the system via the Discover app so I won t go into more details here. However, while Flatpak works great for desktop applications not every piece of software is currently available, and Flatpak is also not designed for other types of programs like system services or command-line tools. Fortunately there are several ways to add software to the Steam Deck without touching the root filesystem, each one with different pros and cons. I will probably talk about some of them in the future, but in this post I m going to focus on one that is already available in the system: systemd-sysext. About systemd-sysext This is a tool included in recent versions of systemd and it is designed to add additional files (in the form of system extensions) to an otherwise immutable root filesystem. Each one of these extensions contains a set of files. When extensions are enabled (aka merged ) those files will appear on the root filesystem using overlayfs. From then on the user can open and run them normally as if they had been installed with a package manager. Merged extensions are seamlessly integrated with the rest of the OS. Since extensions are just collections of files they can be used to add new applications but also other things like system services, development tools, language packs, etc. Creating an extension: yakuake I m using yakuake as an example for this tutorial since the extension is very easy to create, it is an application that some users are demanding and is not easy to distribute with Flatpak. So let s create a yakuake extension. Here are the steps: 1) Create a directory and unpack the files there:
$ mkdir yakuake
$ wget
$ tar -C yakuake -xaf yakuake-*.tar.zst usr
2) Create a file called extension-release.NAME under usr/lib/extension-release.d with the fields ID and VERSION_ID taken from the Steam Deck s /etc/os-release file.
$ mkdir -p yakuake/usr/lib/extension-release.d/
$ echo ID=steamos > yakuake/usr/lib/extension-release.d/extension-release.yakuake
$ echo VERSION_ID=3.3.1 >> yakuake/usr/lib/extension-release.d/extension-release.yakuake
3) Create an image file with the contents of the extension:
$ mksquashfs yakuake yakuake.raw
That s it! The extension is ready. A couple of important things: image files must have the .raw suffix and, despite the name, they can contain any filesystem that the OS can mount. In this example I used SquashFS but other alternatives like EroFS or ext4 are equally valid. NOTE: systemd-sysext can also use extensions from plain directories (i.e skipping the mksquashfs part). Unfortunately we cannot use them in our case because overlayfs does not work with the casefold feature that is enabled on the Steam Deck. Using the extension Once the extension is created you simply need to copy it to a place where systemd-systext can find it. There are several places where they can be installed (see the manual for a list) but due to the Deck s partition layout and the potentially large size of some extensions it probably makes more sense to store them in the home partition and create a link from one of the supported locations (/var/lib/extensions in this example):
(deck@steamdeck ~)$ mkdir extensions
(deck@steamdeck ~)$ scp user@host:/path/to/yakuake.raw extensions/
(deck@steamdeck ~)$ sudo ln -s $PWD/extensions /var/lib/extensions
Once the extension is installed in that directory you only need to enable and start systemd-sysext:
(deck@steamdeck ~)$ sudo systemctl enable systemd-sysext
(deck@steamdeck ~)$ sudo systemctl start systemd-sysext
After this, if everything went fine you should be able to see (and run) /usr/bin/yakuake. The files should remain there from now on, also if you reboot the device. You can see what extensions are enabled with this command:
$ systemd-sysext status
/opt      none       -
/usr      yakuake    Tue 2022-09-13 18:21:53 CEST
If you add or remove extensions from the directory then a simple systemd-sysext refresh is enough to apply the changes. Unfortunately, and unlike distro packages, extensions don t have any kind of post-installation hooks or triggers, so in the case of Yakuake you probably won t see an entry in the KDE application menu immediately after enabling the extension. You can solve that by running kbuildsycoca5 once from the command line. Limitations and caveats Using systemd extensions is generally very easy but there are some things that you need to take into account:
  1. Using extensions is easy (you put them in the directory and voil !). However, creating extensions is not necessarily always easy. To begin with, any libraries, files, etc., that your extensions may need should be either present in the root filesystem or provided by the extension itself. You may need to combine files from different sources or packages into a single extension, or compile them yourself.
  2. In particular, if the extension contains binaries they should probably come from the Steam Deck repository or they should be built to work with those packages. If you need to build your own binaries then having a SteamOS virtual machine can be handy. There you can install all development files and also test that everything works as expected. One could also create a Steam Deck SDK extension with all the necessary files to develop directly on the Deck
  3. Extensions are not distribution packages, they don t have dependency information and therefore they should be self-contained. They also lack triggers and other features available in packages. For desktop applications I still recommend using a system like Flatpak when possible.
  4. Extensions are tied to a particular version of the OS and, as explained above, the ID and VERSION_ID of each extension must match the values from /etc/os-release. If the fields don t match then the extension will be ignored. This is to be expected because there s no guarantee that a particular extension is going to work with a different version of the OS. This can happen after a system update. In the best case one simply needs to update the extension s VERSION_ID, but in some cases it might be necessary to create the extension again with different/updated files.
  5. Extensions only install files in /usr and /opt. Any other file in the image will be ignored. This can be a problem if a particular piece of software needs files in other directories.
  6. When extensions are enabled the /usr and /opt directories become read-only because they are now part of an overlayfs. They will remain read-only even if you run steamos-readonly disable !!. If you really want to make the rootfs read-write you need to disable the extensions (systemd-sysext unmerge) first.
  7. Unlike Flatpak or Podman (including toolbox / distrobox), this is (by design) not meant to isolate the contents of the extension from the rest of the system, so you should be careful with what you re installing. On the other hand, this lack of isolation makes systemd-sysext better suited to some use cases than those container-based systems.
Conclusion systemd extensions are an easy way to add software (or data files) to the immutable OS of the Steam Deck in a way that is seamlessly integrated with the rest of the system. Creating them can be more or less easy depending on the case, but using them is extremely simple. Extensions are not packages, and systemd-sysext is not a package manager or a general-purpose tool to solve all problems, but if you are aware of its limitations it can be a practical tool. It is also possible to share extensions with other users, but here the usual warning against installing binaries from untrusted sources applies. Use with caution, and enjoy!

2 September 2022

John Goerzen: Dead USB Drives Are Fine: Building a Reliable Sneakernet

OK, you re probably thinking. John, you talk a lot about things like Gopher and personal radios, and now you want to talk about building a reliable network out of USB drives? Well, yes. In fact, I ve already done it.

What is sneakernet? Normally, sneakernet is a sort of tongue-in-cheek reference to using disconnected storage to transport data or messages. By disconnect storage I mean anything like CD-ROMs, hard drives, SD cards, USB drives, and so forth. There are times when loading up 12TB on a device and driving it across town is just faster and easier than using the Internet for the same. And, sometimes you need to get data to places that have no Internet at all. Another reason for sneakernet is security. For instance, if your backup system is online, and your systems being backed up are online, then it could become possible for an attacker to destroy both your primary copy of data and your backups. Or, you might use a dedicated computer with no network connection to do GnuPG (GPG) signing.

What about reliable sneakernet, then? TCP is often considered a reliable protocol. That means that the sending side is generally able to tell if its message was properly received. As with most reliable protocols, we have these components:
  1. After transmitting a piece of data, the sender retains it.
  2. After receiving a piece of data, the receiver sends an acknowledgment (ACK) back to the sender.
  3. Upon receiving the acknowledgment, the sender removes its buffered copy of the data.
  4. If no acknowledgment is received at the sender, it retransmits the data, in case it gets lost in transit.
  5. It reorders any packets that arrive out of order, so that the recipient s data stream is ordered correctly.
Now, a lot of the things I just mentioned for sneakernet are legendarily unreliable. USB drives fail, CD-ROMs get scratched, hard drives get banged up. Think about putting these things in a bicycle bag or airline luggage. Some of them are going to fail. You might think, well, I ll just copy files to a USB drive instead of move them, and once I get them onto the destination machine, I ll delete them from the source. Congratulations! You are a human retransmit algorithm! We should be able to automate this! And we can.

Enter NNCP NNCP is one of those things that almost defies explanation. It is a toolkit for building asynchronous networks. It can use as a carrier: a pipe, TCP network connection, a mounted filesystem (specifically intended for cases like this), and much more. It also supports multi-hop asynchronous routing and asynchronous meshing, but these are beyond the scope of this particular article. NNCP s transports that involve live communication between two hops already had all the hallmarks of being reliable; there was a positive ACK and retransmit. As of version 8.7.0, NNCP s ACKs themselves can also be asynchronous meaning that every NNCP transport can now be reliable. Yes, that s right. Your ACKs can flow over tapes and USB drives if you want them to. I use this for archiving and backups. If you aren t already familiar with NNCP, you might take a look at my NNCP page. I also have a lot of blog posts about NNCP. Those pages describe the basics of NNCP: the packet (the unit of transmission in NNCP, which can be tiny or many TB), the end-to-end encryption, and so forth. The new command we will now be interested in is nncp-ack.

The Basic Idea Here are the basic steps to processing this stuff with NNCP:
  1. First, we use nncp-xfer -rx to process incoming packets from the USB (or other media) device. This moves them into the NNCP inbound queue, deleting them from the media device, and verifies the packet integrity.
  2. We use nncp-ack -node $NODE to create ACK packets responding to the packets we just loaded into the rx queue. It writes a list of generated ACKs onto fd 4, which we save off for later use.
  3. We run nncp-toss -seen to process the incoming queue. The use of -seen causes NNCP to remember the hashes of packets seen before, so a duplicate of an already-seen packet will not be processed twice. This command also processes incoming ACKs for packets we ve sent out previously; if they pass verification, the relevant packets are removed from the local machine s tx queue.
  4. Now, we use nncp-xfer -keep -tx -mkdir -node $NODE to send outgoing packets to a given node by writing them to a given directory on the media device. -keep causes them to remain in the outgoing queue.
  5. Finally, we use the list of generated ACK packets saved off in step 2 above. That list is passed to nncp-rm -node $NODE -pkt < $FILE to remove those specific packets from the outbound queue. The reason is that there will never be an ACK of ACK packet (that would create an infinite loop), so if we don t delete them in this manner, they would hang around forever.
You can see these steps follow the same basic outline on upstream s nncp-ack page. One thing to keep in mind: if anything else is running nncp-toss, there is a chance of a race condition between steps 1 and 2 (if nncp-toss gets to it first, it might not get an ack generated). This would sort itself out eventually, presumably, as the sender would retransmit and it would be ACKed later.

Further ideas NNCP guarantees the integrity of packets, but not ordering between packets; if you need that, you might look into my Filespooler program. It is designed to work with NNCP and can provide ordered processing.

An example script Here is a script you might try for this sort of thing. It may have more logic than you need really, you just need the steps above but hopefully it is clear.
set -eo pipefail
# The local node name
NODENAME=" hostname "
# All nodes.  NODENAME should be in this list.
ALLNODES="node1 node2 node3"
# If you need to sudo, use something like RUNNNCP="sudo -Hu nncp"
ACKPATH=" mktemp -d "
# Process incoming packets.
# Parameters: $1 - the path to scan.  Must contain a directory
# named "nncp".
procrxpath ()  
    while [ -n "$1" ]; do
        if ! [ -d "$BASEPATH" ]; then
            echo "$BASEPATH doesn't exist; skipping"
        echo " *** Incoming: processing $BASEPATH"
        TMPDIR=" mktemp -d "
        # This rsync and the one below can help with
        # certain permission issues from weird foreign
        # media.  You could just eliminate it and
        # always use $BASEPATH instead of $TMPDIR below.
        rsync -rt "$BASEPATH/" "$TMPDIR/"
        # You may need these next two lines if using sudo as above.
        # chgrp -R nncp "$TMPDIR"
        # chmod -R g+rwX "$TMPDIR"
        echo "     Running nncp-xfer -rx"
        $RUNNNCP $NNCPPATH/nncp-xfer -progress -rx "$TMPDIR"
        for NODE in $ALLNODES; do
                if [ "$NODE" != "$NODENAME" ]; then
                        echo "     Running nncp-ack for $NODE"
                        # Now, we generate ACK packets for each node we will
                        # process.  nncp-ack writes a list of the created
                        # ACK packets to fd 4.  We'll use them later.
                        # If using sudo, add -C 5 after $RUNNNCP.
                        $RUNNNCP $NNCPPATH/nncp-ack -progress -node "$NODE" \
                           4>> "$ACKPATH/$NODE"
        rsync --delete -rt "$TMPDIR/" "$BASEPATH/"
        rm -fr "$TMPDIR"
proctxpath ()  
    while [ -n "$1" ]; do
        if ! [ -d "$BASEPATH" ]; then
            echo "$BASEPATH doesn't exist; skipping"
        echo " *** Outgoing: processing $BASEPATH"
        TMPDIR=" mktemp -d "
        rsync -rt "$BASEPATH/" "$TMPDIR/"
        # You may need these two lines if using sudo:
        # chgrp -R nncp "$TMPDIR"
        # chmod -R g+rwX "$TMPDIR"
        for DESTHOST in $ALLNODES; do
            if [ "$DESTHOST" = "$NODENAME" ]; then
            # Copy outgoing packets to this node, but keep them in the outgoing
            # queue with -keep.
            $RUNNNCP $NNCPPATH/nncp-xfer -keep -tx -mkdir -node "$DESTHOST" -progress "$TMPDIR"
            # Here is the key: that list of ACK packets we made above - now we delete them.
            # There will never be an ACK for an ACK, so they'd keep sending forever
            # if we didn't do this.
            if [ -f "$ACKPATH/$DESTHOST" ]; then
                echo "nncp-rm for node $DESTHOST"
                $RUNNNCP $NNCPPATH/nncp-rm -debug -node "$DESTHOST" -pkt < "$ACKPATH/$DESTHOST"
        rsync --delete -rt "$TMPDIR/" "$BASEPATH/"
        rm -rf "$TMPDIR"
        # We only want to write stuff once.
        return 0
procrxpath "$MEDIABASE"/*
echo " *** Initial tossing..."
# We make sure to use -seen to rule out duplicates.
$RUNNNCP $NNCPPATH/nncp-toss -progress -seen
proctxpath "$MEDIABASE"/*
echo "You can unmount devices now."
echo "Done."

This post is also available on my webiste, where it may be periodically updated.

1 August 2022

Sergio Talens-Oliag: Using Git Server Hooks on GitLab CE to Validate Tags

Since a long time ago I ve been a gitlab-ce user, in fact I ve set it up on three of the last four companies I ve worked for (initially I installed it using the omnibus packages on a debian server but on the last two places I moved to the docker based installation, as it is easy to maintain and we don t need a big installation as the teams using it are small). On the company I work for now (kyso) we are using it to host all our internal repositories and to do all the CI/CD work (the automatic deployments are triggered by web hooks in some cases, but the rest is all done using gitlab-ci). The majority of projects are using nodejs as programming language and we have automated the publication of npm packages on our gitlab instance npm registry and even the publication into the npmjs registry. To publish the packages we have added rules to the gitlab-ci configuration of the relevant repositories and we publish them when a tag is created. As the we are lazy by definition, I configured the system to use the tag as the package version; I tested if the contents of the package.json where in sync with the expected version and if it was not I updated it and did a force push of the tag with the updated file using the following code on the script that publishes the package:
# Update package version & add it to the .build-args
INITIAL_PACKAGE_VERSION="$(npm pkg get version tr -d '"')"
npm version --allow-same --no-commit-hooks --no-git-tag-version \
UPDATED_PACKAGE_VERSION="$(npm pkg get version tr -d '"')"
# Update tag if the version was updated or abort
  if [ -n "$CI_GIT_USER" ] && [ -n "$CI_GIT_TOKEN" ]; then
    git commit -m "Updated version from tag $CI_COMMIT_TAG" package.json
    git tag -f "$CI_COMMIT_TAG" -m "Updated version from tag"
    git push -f -o ci.skip origin "$CI_COMMIT_TAG"
    echo "!!! ERROR !!!"
    echo "The updated tag could not be uploaded."
    echo "Set CI_GIT_USER and CI_GIT_TOKEN or fix the 'package.json' file"
    echo "!!! ERROR !!!"
    exit 1
This feels a little dirty (we are leaving commits on the tag but not updating the original branch); I thought about trying to find the branch using the tag and update it, but I drop the idea pretty soon as there were multiple issues to consider (i.e. we can have tags pointing to commits present in multiple branches and even if it only points to one the tag does not have to be the HEAD of the branch making the inclusion difficult). In any case this system was working, so we left it until we started to publish to the NPM Registry; as we are using a token to push the packages that we don t want all developers to have access to (right now it would not matter, but when the team grows it will) I started to use gitlab protected branches on the projects that need it and adjusting the .npmrc file using protected variables. The problem then was that we can no longer do a standard force push for a branch (that is the main point of the protected branches feature) unless we use the gitlab api, so the tags with the wrong version started to fail. As the way things were being done seemed dirty anyway I thought that the best way of fixing things was to forbid users to push a tag that includes a version that does not match the package.json version. After thinking about it we decided to use githooks on the gitlab server for the repositories that need it, as we are only interested in tags we are going to use the update hook; it is executed once for each ref to be updated, and takes three parameters:
  • the name of the ref being updated,
  • the old object name stored in the ref,
  • and the new object name to be stored in the ref.
To install our hook we have found the gitaly relative path of each repo and located it on the server filesystem (as I said we are using docker and the gitlab s data directory is on /srv/gitlab/data, so the path to the repo has the form /srv/gitlab/data/git-data/repositories/@hashed/xx/yy/hash.git). Once we have the directory we need to:
  • create a custom_hooks sub directory inside it,
  • add the update script (as we only need one script we used that instead of creating an update.d directory, the good thing is that this will also work with a standard git server renaming the base directory to hooks instead of custom_hooks),
  • make it executable, and
  • change the directory and file ownership to make sure it can be read and executed from the gitlab container
On a console session:
$ cd /srv/gitlab/data/git-data/repositories/@hashed/xx/yy/hash.git
$ mkdir custom_hooks
$ edit_or_copy custom_hooks/update
$ chmod 0755 custom_hooks/update
$ chown --reference=. -R custom_hooks
The update script we are using is as follows:
set -e
# kyso update hook
# Right now it checks version.txt or package.json versions against the tag name
# (it supports a 'v' prefix on the tag)
# Arguments
# Initial test
if [ -z "$ref_name" ]    [ -z "$old_rev" ]   [ -z "$new_rev" ]; then
  echo "usage: $0 <ref> <oldrev> <newrev>" >&2
  exit 1
# Get the tag short name
tag_name="$ ref_name##refs/tags/ "
# Exit if the update is not for a tag
if [ "$tag_name" = "$ref_name" ]; then
  exit 0
# Get the null rev value (string of zeros)
zero=$(git hash-object --stdin </dev/null   tr '0-9a-f' '0')
# Get if the tag is new or not
if [ "$old_rev" = "$zero" ]; then
# Get the type of revision:
# - delete: if the new_rev is zero
# - commit: annotated tag
# - tag: un-annotated tag
if [ "$new_rev" = "$zero" ]; then
  new_rev_type="$(git cat-file -t "$new_rev")"
# Exit if we are deleting a tag (nothing to check here)
if [ "$new_rev_type" = "delete" ]; then
  exit 0
# Check the version against the tag (supports version.txt & package.json)
if git cat-file -e "$new_rev:version.txt" >/dev/null 2>&1; then
  version="$(git cat-file -p "$new_rev:version.txt")"
  if [ "$version" = "$tag_name" ]   [ "$version" = "$ tag_name#v " ]; then
    exit 0
    EMSG="tag '$tag_name' and 'version.txt' contents '$version' don't match"
    echo "GL-HOOK-ERR: $EMSG"
    exit 1
elif git cat-file -e "$new_rev:package.json" >/dev/null 2>&1; then
    git cat-file -p "$new_rev:package.json"   jsonpath version   tr -d '\[\]"'
  if [ "$version" = "$tag_name" ]   [ "$version" = "$ tag_name#v " ]; then
    exit 0
    EMSG="tag '$tag_name' and 'package.json' version '$version' don't match"
    echo "GL-HOOK-ERR: $EMSG"
    exit 1
  # No version.txt or package.json file found
  exit 0
Some comments about it:
  • we are only looking for tags, if the ref_name does not have the prefix refs/tags/ the script does an exit 0,
  • although we are checking if the tag is new or not we are not using the value (in gitlab that is handled by the protected tag feature),
  • if we are deleting a tag the script does an exit 0, we don t need to check anything in that case,
  • we are ignoring if the tag is annotated or not (we set the new_rev_type to tag or commit, but we don t use the value),
  • we test first the version.txt file and if it does not exist we check the package.json file, if it does not exist either we do an exit 0, as there is no version to check against and we allow that on a tag,
  • we add the GL-HOOK-ERR: prefix to the messages to show them on the gitlab web interface (can be tested creating a tag from it),
  • to get the version on the package.json file we use the jsonpath binary (it is installed by the jsonpath ruby gem) because it is available on the gitlab container (initially I used sed to get the value, but a real JSON parser is always a better option).
Once the hook is installed when a user tries to push a tag to a repository that has a version.txt file or package.json file and the tag does not match the version (if version.txt is present it takes precedence) the push fails. If the tag matches or the files are not present the tag is added if the user has permission to add it in gitlab (our hook is only executed if the user is allowed to create or update the tag).

17 July 2022

Martin-&#201;ric Racine: Trying to chainload iPXE on old Etherboot hardware

Among my collection of PC hardware, I have a few rarities whose netboot implementation predates PXE. Since I recently managed to configure dnsmasq as a potent TFTP and PXE server, I figured that I'd try chainloading iPXE via BOOTP options. This required preparing a boot image using antiquated tools: $ sudo mkelf-linux --param=autoboot --output=/srv/tftp/ipxe.nbi /srv/tftp/ipxe.lkrn The host succesufully loads the boot image, except that the iPXE blob fails to find the network card:
Any ideas?

7 July 2022

Jonathan Dowland: Musick To Play In The Dark 2

This took a while to arrive! After the success of the reissue of Coil's Musick To Play In The Dark, it was inevitable that the second edition would also be reissued. The pre-order opened late November 2021 and mine arrived in late April this year.
Record cover
I was toying with the idea of ordering one of the most exclusive editions direct from DAIS Records, in particular the glow in the dark one, but with international shipping the cost was pretty high. I went with a UK distributor (Boomkat) instead, who had their own exclusive edition: black-in-purple-in-clear.
I'm happy with my decision: it's one of the most interesting variants I own, and from what I've read, glow in the dark vinyl never sounds great anyway. (I think I have at least one glow in the dark 7" single somewhere) Stand-out track: Tiny Golden Books Since I didn't say so last time, the stand-out track on volume one is Red Birds Will Fly Out of the East and Destroy Paris in a Night, but both volumes are full of really interesting moments ( One day, your eggs are going to hatch and some very strange birds are going to emerge. )

24 June 2022

Kees Cook: finding binary differences

As part of the continuing work to replace 1-element arrays in the Linux kernel, it s very handy to show that a source change has had no executable code difference. For example, if you started with this:
struct foo  
    unsigned long flags;
    u32 length;
    u32 data[1];
void foo_init(int count)
    struct foo *instance;
    size_t bytes = sizeof(*instance) + sizeof(u32) * (count - 1);
    instance = kmalloc(bytes, GFP_KERNEL);
And you changed only the struct definition:
-    u32 data[1];
+    u32 data[];
The bytes calculation is going to be incorrect, since it is still subtracting 1 element s worth of space from the desired count. (And let s ignore for the moment the open-coded calculation that may end up with an arithmetic over/underflow here; that can be solved separately by using the struct_size() helper or the size_mul(), size_add(), etc family of helpers.) The missed adjustment to the size calculation is relatively easy to find in this example, but sometimes it s much less obvious how structure sizes might be woven into the code. I ve been checking for issues by using the fantastic diffoscope tool. It can produce a LOT of noise if you try to compare builds without keeping in mind the issues solved by reproducible builds, with some additional notes. I prepare my build with the known to disrupt code layout options disabled, but with debug info enabled:
$ OUT=gcc
$ make $KBF O=$OUT allmodconfig
$ ./scripts/config --file $OUT/.config \
$ make $KBF O=$OUT olddefconfig
Then I build a stock target, saving the output in before . In this case, I m examining drivers/scsi/megaraid/:
$ make -jN $KBF O=$OUT drivers/scsi/megaraid/
$ mkdir -p $OUT/before
$ cp $OUT/drivers/scsi/megaraid/*.o $OUT/before/
Then I patch and build a modified target, saving the output in after :
$ vi the/source/code.c
$ make -jN $KBF O=$OUT drivers/scsi/megaraid/
$ mkdir -p $OUT/after
$ cp $OUT/drivers/scsi/megaraid/*.o $OUT/after/
And then run diffoscope:
$ diffoscope $OUT/before/ $OUT/after/
If diffoscope output reports nothing, then we re done. Usually, though, when source lines move around other stuff will shift too (e.g. WARN macros rely on line numbers, so the bug table may change contents a bit, etc), and diffoscope output will look noisy. To examine just the executable code, the command that diffoscope used is reported in the output, and we can run it directly, but with possibly shifted line numbers not reported. i.e. running objdump without --line-numbers:
$ ARGS="--disassemble --demangle --reloc --no-show-raw-insn --section=.text"
$ for i in $(cd $OUT/before && echo *.o); do
        echo $i
        diff -u <(objdump $ARGS $OUT/before/$i   sed "0,/^Disassembly/d") \
                <(objdump $ARGS $OUT/after/$i    sed "0,/^Disassembly/d")
If I see an unexpected difference, for example:
-    c120:      movq   $0x0,0x800(%rbx)
+    c120:      movq   $0x0,0x7f8(%rbx)
Then I'll search for the pattern with line numbers added to the objdump output:
$ vi <(objdump --line-numbers $ARGS $OUT/after/megaraid_sas_fp.o)
I'd search for "0x0,0x7f8", find the source file and line number above it, open that source file at that position, and look to see where something was being miscalculated:
$ vi drivers/scsi/megaraid/megaraid_sas_fp.c +329
Once tracked down, I'd start over at the "patch and build a modified target" step above, repeating until there were no differences. For example, in the starting example, I'd also need to make this change:
-    size_t bytes = sizeof(*instance) + sizeof(u32) * (count - 1);
+    size_t bytes = sizeof(*instance) + sizeof(u32) * count;
Though, as hinted earlier, better yet would be:
-    size_t bytes = sizeof(*instance) + sizeof(u32) * (count - 1);
+    size_t bytes = struct_size(instance, data, count);
But sometimes adding the helper usage will add binary output differences since they're performing overflow checking that might saturate at SIZE_MAX. To help with patch clarity, those changes can be done separately from fixing the array declaration.

2022, Kees Cook. This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 License.
CC BY-SA 4.0

18 June 2022

Bastian Venthur: blag is now available in Debian

Last year, I wrote my own blog-aware static site generator in Python. I called it blag named after the blag of the webcomic xkcd. Now I finally got around packaging- and uploading blag to Debian. It passed the NEW queue and is now part of the distribution. That means if you re using Debian, you can install it via:
sudo aptitude install blag
Ubuntu will probably follow soon. For every other system, blag is also available on PyPI:
pip install blag
To get started, you can
mkdir blog && cd blog
blag quickstart                        # fill out some info
nvim content/            # write some content
blag build                             # build the website
Blag is aware of articles and pages: the difference is that articles are part of the blog and will be added to the atom feed, the archive and aggregated in the tag pages. Pages are just rendered out to HTML. Articles and pages can be freely mixed in the content directory, what differentiates an article from a page is the existence of the dade metadata element:
title: My first article
description: Short description of the article
date: 2022-06-18 23:00
tags: blogging, markdown
## Hello World!
Lorem ipsum.
blag also comes with a dev-server that rebuilds the website automatically on every change detected, you can start it using:
blag serve
The default theme looks quite ugly, and you probably want to create your own styling to make it more beautiful. The process is not very difficult if you re familiar with jinja templating. Help on that can be found in the Templating section of the online documentation, the offline version in the blag-doc package, or the man page, respectively. Speaking of the blag-doc package: packaging it was surprisingly tricky, and it also took me a lot of trial and error to realize that dh_sphinxdocs alone does not automatically put the generated html output into the appropriate package, you rather have to list them in the (i.e. so dh_installdocs can install them properly.

17 June 2022

Antoine Beaupr : Matrix notes

I have some concerns about Matrix (the protocol, not the movie that came out recently, although I do have concerns about that as well). I've been watching the project for a long time, and it seems more a promising alternative to many protocols like IRC, XMPP, and Signal. This review may sound a bit negative, because it focuses on those concerns. I am the operator of an IRC network and people keep asking me to bridge it with Matrix. I have myself considered just giving up on IRC and converting to Matrix. This space is a living document exploring my research of that problem space. The TL;DR: is that no, I'm not setting up a bridge just yet, and I'm still on IRC. This article was written over the course of the last three months, but I have been watching the Matrix project for years (my logs seem to say 2016 at least). The article is rather long. It will likely take you half an hour to read, so copy this over to your ebook reader, your tablet, or dead trees, and lean back and relax as I show you around the Matrix. Or, alternatively, just jump to a section that interest you, most likely the conclusion.

Introduction to Matrix Matrix is an "open standard for interoperable, decentralised, real-time communication over IP. It can be used to power Instant Messaging, VoIP/WebRTC signalling, Internet of Things communication - or anywhere you need a standard HTTP API for publishing and subscribing to data whilst tracking the conversation history". It's also (when compared with XMPP) "an eventually consistent global JSON database with an HTTP API and pubsub semantics - whilst XMPP can be thought of as a message passing protocol." According to their FAQ, the project started in 2014, has about 20,000 servers, and millions of users. Matrix works over HTTPS but over a special port: 8448.

Security and privacy I have some concerns about the security promises of Matrix. It's advertised as a "secure" with "E2E [end-to-end] encryption", but how does it actually work?

Data retention defaults One of my main concerns with Matrix is data retention, which is a key part of security in a threat model where (for example) an hostile state actor wants to surveil your communications and can seize your devices. On IRC, servers don't actually keep messages all that long: they pass them along to other servers and clients as fast as they can, only keep them in memory, and move on to the next message. There are no concerns about data retention on messages (and their metadata) other than the network layer. (I'm ignoring the issues with user registration, which is a separate, if valid, concern.) Obviously, an hostile server could log everything passing through it, but IRC federations are normally tightly controlled. So, if you trust your IRC operators, you should be fairly safe. Obviously, clients can (and often do, even if OTR is configured!) log all messages, but this is generally not the default. Irssi, for example, does not log by default. IRC bouncers are more likely to log to disk, of course, to be able to do what they do. Compare this to Matrix: when you send a message to a Matrix homeserver, that server first stores it in its internal SQL database. Then it will transmit that message to all clients connected to that server and room, and to all other servers that have clients connected to that room. Those remote servers, in turn, will keep a copy of that message and all its metadata in their own database, by default forever. On encrypted rooms those messages are encrypted, but not their metadata. There is a mechanism to expire entries in Synapse, but it is not enabled by default. So one should generally assume that a message sent on Matrix is never expired.

GDPR in the federation But even if that setting was enabled by default, how do you control it? This is a fundamental problem of the federation: if any user is allowed to join a room (which is the default), those user's servers will log all content and metadata from that room. That includes private, one-on-one conversations, since those are essentially rooms as well. In the context of the GDPR, this is really tricky: who is the responsible party (known as the "data controller") here? It's basically any yahoo who fires up a home server and joins a room. In a federated network, one has to wonder whether GDPR enforcement is even possible at all. But in Matrix in particular, if you want to enforce your right to be forgotten in a given room, you would have to:
  1. enumerate all the users that ever joined the room while you were there
  2. discover all their home servers
  3. start a GDPR procedure against all those servers
I recognize this is a hard problem to solve while still keeping an open ecosystem. But I believe that Matrix should have much stricter defaults towards data retention than right now. Message expiry should be enforced by default, for example. (Note that there are also redaction policies that could be used to implement part of the GDPR automatically, see the privacy policy discussion below on that.) Also keep in mind that, in the brave new peer-to-peer world that Matrix is heading towards, the boundary between server and client is likely to be fuzzier, which would make applying the GDPR even more difficult. Update: this comment links to this post (in german) which apparently studied the question and concluded that Matrix is not GDPR-compliant. In fact, maybe Synapse should be designed so that there's no configurable flag to turn off data retention. A bit like how most system loggers in UNIX (e.g. syslog) come with a log retention system that typically rotate logs after a few weeks or month. Historically, this was designed to keep hard drives from filling up, but it also has the added benefit of limiting the amount of personal information kept on disk in this modern day. (Arguably, syslog doesn't rotate logs on its own, but, say, Debian GNU/Linux, as an installed system, does have log retention policies well defined for installed packages, and those can be discussed. And "no expiry" is definitely a bug. privacy policy When I first looked at Matrix, five years ago, was called and had a rather dubious privacy policy:
We currently use cookies to support our use of Google Analytics on the Website and Service. Google Analytics collects information about how you use the Website and Service. [...] This helps us to provide you with a good experience when you browse our Website and use our Service and also allows us to improve our Website and our Service.
When I asked Matrix people about why they were using Google Analytics, they explained this was for development purposes and they were aiming for velocity at the time, not privacy (paraphrasing here). They also included a "free to snitch" clause:
If we are or believe that we are under a duty to disclose or share your personal data, we will do so in order to comply with any legal obligation, the instructions or requests of a governmental authority or regulator, including those outside of the UK.
Those are really broad terms, above and beyond what is typically expected legally. Like the current retention policies, such user tracking and ... "liberal" collaboration practices with the state set a bad precedent for other home servers. Thankfully, since the above policy was published (2017), the GDPR was "implemented" (2018) and it seems like both the privacy policy and the privacy policy have been somewhat improved since. Notable points of the new privacy policies:
  • the "federation" section actually outlines that "Federated homeservers and Matrix clients which respect the Matrix protocol are expected to honour these controls and redaction/erasure requests, but other federated homeservers are outside of the span of control of Element, and we cannot guarantee how this data will be processed"
  • 2.6: users under the age of 16 should not use the service
  • 2.10: Upcloud, Mythic Beast, Amazon, and CloudFlare possibly have access to your data (it's nice to at least mention this in the privacy policy: many providers don't even bother admitting to this kind of delegation)
  • Element 2.2.1: mentions many more third parties (Twilio, Stripe, Quaderno, LinkedIn, Twitter, Google, Outplay, PipeDrive, HubSpot, Posthog, Sentry, and Matomo (phew!) used when you are paying for hosting
I'm not super happy with all the trackers they have on the Element platform, but then again you don't have to use that service. Your favorite homeserver (assuming you are not on probably has their own Element deployment, hopefully without all that garbage. Overall, this is all a huge improvement over the previous privacy policy, so hats off to the Matrix people for figuring out a reasonable policy in such a tricky context. I particularly like this bit:
We will forget your copy of your data upon your request. We will also forward your request to be forgotten onto federated homeservers. However - these homeservers are outside our span of control, so we cannot guarantee they will forget your data.
It's great they implemented those mechanisms and, after all, if there's an hostile party in there, nothing can prevent them from using screenshots to just exfiltrate your data away from the client side anyways, even with services typically seen as more secure, like Signal. As an aside, I also appreciate that has a fairly decent code of conduct, based on the TODO CoC which checks all the boxes in the geekfeminism wiki.

Metadata handling Overall, privacy protections in Matrix mostly concern message contents, not metadata. In other words, who's talking with who, when and from where is not well protected. Compared to a tool like Signal, which goes through great lengths to anonymize that data with features like private contact discovery, disappearing messages, sealed senders, and private groups, Matrix is definitely behind. (Note: there is an issue open about message lifetimes in Element since 2020, but it's not at even at the MSC stage yet.) This is a known issue (opened in 2019) in Synapse, but this is not just an implementation issue, it's a flaw in the protocol itself. Home servers keep join/leave of all rooms, which gives clear text information about who is talking to. Synapse logs may also contain privately identifiable information that home server admins might not be aware of in the first place. Those log rotation policies are separate from the server-level retention policy, which may be confusing for a novice sysadmin. Combine this with the federation: even if you trust your home server to do the right thing, the second you join a public room with third-party home servers, those ideas kind of get thrown out because those servers can do whatever they want with that information. Again, a problem that is hard to solve in any federation. To be fair, IRC doesn't have a great story here either: any client knows not only who's talking to who in a room, but also typically their client IP address. Servers can (and often do) obfuscate this, but often that obfuscation is trivial to reverse. Some servers do provide "cloaks" (sometimes automatically), but that's kind of a "slap-on" solution that actually moves the problem elsewhere: now the server knows a little more about the user. Overall, I would worry much more about a Matrix home server seizure than a IRC or Signal server seizure. Signal does get subpoenas, and they can only give out a tiny bit of information about their users: their phone number, and their registration, and last connection date. Matrix carries a lot more information in its database.

Amplification attacks on URL previews I (still!) run an Icecast server and sometimes share links to it on IRC which, obviously, also ends up on (more than one!) Matrix home servers because some people connect to IRC using Matrix. This, in turn, means that Matrix will connect to that URL to generate a link preview. I feel this outlines a security issue, especially because those sockets would be kept open seemingly forever. I tried to warn the Matrix security team but somehow, I don't think this issue was taken very seriously. Here's the disclosure timeline:
  • January 18: contacted Matrix security
  • January 19: response: already reported as a bug
  • January 20: response: can't reproduce
  • January 31: timeout added, considered solved
  • January 31: I respond that I believe the security issue is underestimated, ask for clearance to disclose
  • February 1: response: asking for two weeks delay after the next release (1.53.0) including another patch, presumably in two weeks' time
  • February 22: Matrix 1.53.0 released
  • April 14: I notice the release, ask for clearance again
  • April 14: response: referred to the public disclosure
There are a couple of problems here:
  1. the bug was publicly disclosed in September 2020, and not considered a security issue until I notified them, and even then, I had to insist
  2. no clear disclosure policy timeline was proposed or seems established in the project (there is a security disclosure policy but it doesn't include any predefined timeline)
  3. I wasn't informed of the disclosure
  4. the actual solution is a size limit (10MB, already implemented), a time limit (30 seconds, implemented in PR 11784), and a content type allow list (HTML, "media" or JSON, implemented in PR 11936), and I'm not sure it's adequate
  5. (pure vanity:) I did not make it to their Hall of fame
I'm not sure those solutions are adequate because they all seem to assume a single home server will pull that one URL for a little while then stop. But in a federated network, many (possibly thousands) home servers may be connected in a single room at once. If an attacker drops a link into such a room, all those servers would connect to that link all at once. This is an amplification attack: a small amount of traffic will generate a lot more traffic to a single target. It doesn't matter there are size or time limits: the amplification is what matters here. It should also be noted that clients that generate link previews have more amplification because they are more numerous than servers. And of course, the default Matrix client (Element) does generate link previews as well. That said, this is possibly not a problem specific to Matrix: any federated service that generates link previews may suffer from this. I'm honestly not sure what the solution is here. Maybe moderation? Maybe link previews are just evil? All I know is there was this weird bug in my Icecast server and I tried to ring the bell about it, and it feels it was swept under the rug. Somehow I feel this is bound to blow up again in the future, even with the current mitigation.

Moderation In Matrix like elsewhere, Moderation is a hard problem. There is a detailed moderation guide and much of this problem space is actively worked on in Matrix right now. A fundamental problem with moderating a federated space is that a user banned from a room can rejoin the room from another server. This is why spam is such a problem in Email, and why IRC networks have stopped federating ages ago (see the IRC history for that fascinating story).

The mjolnir bot The mjolnir moderation bot is designed to help with some of those things. It can kick and ban users, redact all of a user's message (as opposed to one by one), all of this across multiple rooms. It can also subscribe to a federated block list published by to block known abusers (users or servers). Bans are pretty flexible and can operate at the user, room, or server level. Matrix people suggest making the bot admin of your channels, because you can't take back admin from a user once given.

The command-line tool There's also a new command line tool designed to do things like:
  • System notify users (all users/users from a list, specific user)
  • delete sessions/devices not seen for X days
  • purge the remote media cache
  • select rooms with various criteria (external/local/empty/created by/encrypted/cleartext)
  • purge history of theses rooms
  • shutdown rooms
This tool and Mjolnir are based on the admin API built into Synapse.

Rate limiting Synapse has pretty good built-in rate-limiting which blocks repeated login, registration, joining, or messaging attempts. It may also end up throttling servers on the federation based on those settings.

Fundamental federation problems Because users joining a room may come from another server, room moderators are at the mercy of the registration and moderation policies of those servers. Matrix is like IRC's +R mode ("only registered users can join") by default, except that anyone can register their own homeserver, which makes this limited. Server admins can block IP addresses and home servers, but those tools are not easily available to room admins. There is an API ( in /devtools) but it is not reliable (thanks Austin Huang for the clarification). Matrix has the concept of guest accounts, but it is not used very much, and virtually no client or homeserver supports it. This contrasts with the way IRC works: by default, anyone can join an IRC network even without authentication. Some channels require registration, but in general you are free to join and look around (until you get blocked, of course). I have seen anecdotal evidence (CW: Twitter, nitter link) that "moderating bridges is hell", and I can imagine why. Moderation is already hard enough on one federation, when you bridge a room with another network, you inherit all the problems from that network but without the entire abuse control tools from the original network's API...

Room admins Matrix, in particular, has the problem that room administrators (which have the power to redact messages, ban users, and promote other users) are bound to their Matrix ID which is, in turn, bound to their home servers. This implies that a home server administrators could (1) impersonate a given user and (2) use that to hijack the room. So in practice, the home server is the trust anchor for rooms, not the user themselves. That said, if server B administrator hijack user joe on server B, they will hijack that room on that specific server. This will not (necessarily) affect users on the other servers, as servers could refuse parts of the updates or ban the compromised account (or server). It does seem like a major flaw that room credentials are bound to Matrix identifiers, as opposed to the E2E encryption credentials. In an encrypted room even with fully verified members, a compromised or hostile home server can still take over the room by impersonating an admin. That admin (or even a newly minted user) can then send events or listen on the conversations. This is even more frustrating when you consider that Matrix events are actually signed and therefore have some authentication attached to them, acting like some sort of Merkle tree (as it contains a link to previous events). That signature, however, is made from the homeserver PKI keys, not the client's E2E keys, which makes E2E feel like it has been "bolted on" later.

Availability While Matrix has a strong advantage over Signal in that it's decentralized (so anyone can run their own homeserver,), I couldn't find an easy way to run a "multi-primary" setup, or even a "redundant" setup (even if with a single primary backend), short of going full-on "replicate PostgreSQL and Redis data", which is not typically for the faint of heart.

How this works in IRC On IRC, it's quite easy to setup redundant nodes. All you need is:
  1. a new machine (with it's own public address with an open port)
  2. a shared secret (or certificate) between that machine and an existing one on the network
  3. a connect block on both servers
That's it: the node will join the network and people can connect to it as usual and share the same user/namespace as the rest of the network. The servers take care of synchronizing state: you do not need to worry about replicating a database server. (Now, experienced IRC people will know there's a catch here: IRC doesn't have authentication built in, and relies on "services" which are basically bots that authenticate users (I'm simplifying, don't nitpick). If that service goes down, the network still works, but then people can't authenticate, and they can start doing nasty things like steal people's identity if they get knocked offline. But still: basic functionality still works: you can talk in rooms and with users that are on the reachable network.)

User identities Matrix is more complicated. Each "home server" has its own identity namespace: a specific user (say is bound to that specific home server. If that server goes down, that user is completely disconnected. They could register a new account elsewhere and reconnect, but then they basically lose all their configuration: contacts, joined channels are all lost. (Also notice how the Matrix IDs don't look like a typical user address like an email in XMPP. They at least did their homework and got the allocation for the scheme.)

Rooms Users talk to each other in "rooms", even in one-to-one communications. (Rooms are also used for other things like "spaces", they're basically used for everything, think "everything is a file" kind of tool.) For rooms, home servers act more like IRC nodes in that they keep a local state of the chat room and synchronize it with other servers. Users can keep talking inside a room if the server that originally hosts the room goes down. Rooms can have a local, server-specific "alias" so that, say, is also visible as on the home server. Both addresses refer to the same room underlying room. (Finding this in the Element settings is not obvious though, because that "alias" are actually called a "local address" there. So to create such an alias (in Element), you need to go in the room settings' "General" section, "Show more" in "Local address", then add the alias name (e.g. foo), and then that room will be available on your homeserver as So a room doesn't belong to a server, it belongs to the federation, and anyone can join the room from any serer (if the room is public, or if invited otherwise). You can create a room on server A and when a user from server B joins, the room will be replicated on server B as well. If server A fails, server B will keep relaying traffic to connected users and servers. A room is therefore not fundamentally addressed with the above alias, instead ,it has a internal Matrix ID, which basically a random string. It has a server name attached to it, but that was made just to avoid collisions. That can get a little confusing. For example, the room is an alias on the server, but the room ID is ! That's because the room was created on, but the preferred branding is now. As an aside, rooms, by default, live forever, even after the last user quits. There's an admin API to delete rooms and a tombstone event to redirect to another one, but neither have a GUI yet. The latter is part of MSC1501 ("Room version upgrades") which allows a room admin to close a room, with a message and a pointer to another room.

Spaces Discovering rooms can be tricky: there is a per-server room directory, but people are trying to deprecate it in favor of "Spaces". Room directories were ripe for abuse: anyone can create a room, so anyone can show up in there. It's possible to restrict who can add aliases, but anyways directories were seen as too limited. In contrast, a "Space" is basically a room that's an index of other rooms (including other spaces), so existing moderation and administration mechanism that work in rooms can (somewhat) work in spaces as well. This enables a room directory that works across federation, regardless on which server they were originally created. New users can be added to a space or room automatically in Synapse. (Existing users can be told about the space with a server notice.) This gives admins a way to pre-populate a list of rooms on a server, which is useful to build clusters of related home servers, providing some sort of redundancy, at the room -- not user -- level.

Home servers So while you can workaround a home server going down at the room level, there's no such thing at the home server level, for user identities. So if you want those identities to be stable in the long term, you need to think about high availability. One limitation is that the domain name (e.g. must never change in the future, as renaming home servers is not supported. The documentation used to say you could "run a hot spare" but that has been removed. Last I heard, it was not possible to run a high-availability setup where multiple, separate locations could replace each other automatically. You can have high performance setups where the load gets distributed among workers, but those are based on a shared database (Redis and PostgreSQL) backend. So my guess is it would be possible to create a "warm" spare server of a matrix home server with regular PostgreSQL replication, but that is not documented in the Synapse manual. This sort of setup would also not be useful to deal with networking issues or denial of service attacks, as you will not be able to spread the load over multiple network locations easily. Redis and PostgreSQL heroes are welcome to provide their multi-primary solution in the comments. In the meantime, I'll just point out this is a solution that's handled somewhat more gracefully in IRC, by having the possibility of delegating the authentication layer.

Delegations If you do not want to run a Matrix server yourself, it's possible to delegate the entire thing to another server. There's a server discovery API which uses the .well-known pattern (or SRV records, but that's "not recommended" and a bit confusing) to delegate that service to another server. Be warned that the server still needs to be explicitly configured for your domain. You can't just put:
  "m.server": ""  
... on and start using as a Matrix ID. That's because Matrix doesn't support "virtual hosting" and you'd still be connecting to rooms and people with your identity, not as you would normally expect. This is also why you cannot rename your home server. The server discovery API is what allows servers to find each other. Clients, on the other hand, use the client-server discovery API: this is what allows a given client to find your home server when you type your Matrix ID on login.

Performance The high availability discussion brushed over the performance of Matrix itself, but let's now dig into that.

Horizontal scalability There were serious scalability issues of the main Matrix server, Synapse, in the past. So the Matrix team has been working hard to improve its design. Since Synapse 1.22 the home server can horizontally scale to multiple workers (see this blog post for details) which can make it easier to scale large servers.

Other implementations There are other promising home servers implementations from a performance standpoint (dendrite, Golang, entered beta in late 2020; conduit, Rust, beta; others), but none of those are feature-complete so there's a trade-off to be made there. Synapse is also adding a lot of feature fast, so it's an open question whether the others will ever catch up. (I have heard that Dendrite might actually surpass Synapse in features within a few years, which would put Synapse in a more "LTS" situation.)

Latency Matrix can feel slow sometimes. For example, joining the "Matrix HQ" room in Element (from takes a few minutes and then fails. That is because the home server has to sync the entire room state when you join the room. There was promising work on this announced in the lengthy 2021 retrospective, and some of that work landed (partial sync) in the 1.53 release already. Other improvements coming include sliding sync, lazy loading over federation, and fast room joins. So that's actually something that could be fixed in the fairly short term. But in general, communication in Matrix doesn't feel as "snappy" as on IRC or even Signal. It's hard to quantify this without instrumenting a full latency test bed (for example the tools I used in the terminal emulators latency tests), but even just typing in a web browser feels slower than typing in a xterm or Emacs for me. Even in conversations, I "feel" people don't immediately respond as fast. In fact, this could be an interesting double-blind experiment to make: have people guess whether they are talking to a person on Matrix, XMPP, or IRC, for example. My theory would be that people could notice that Matrix users are slower, if only because of the TCP round-trip time each message has to take.

Transport Some courageous person actually made some tests of various messaging platforms on a congested network. His evaluation was basically:
  • Briar: uses Tor, so unusable except locally
  • Matrix: "struggled to send and receive messages", joining a room takes forever as it has to sync all history, "took 20-30 seconds for my messages to be sent and another 20 seconds for further responses"
  • XMPP: "worked in real-time, full encryption, with nearly zero lag"
So that was interesting. I suspect IRC would have also fared better, but that's just a feeling. Other improvements to the transport layer include support for websocket and the CoAP proxy work from 2019 (targeting 100bps links), but both seem stalled at the time of writing. The Matrix people have also announced the pinecone p2p overlay network which aims at solving large, internet-scale routing problems. See also this talk at FOSDEM 2022.


Onboarding and workflow The workflow for joining a room, when you use Element web, is not great:
  1. click on a link in a web browser
  2. land on (say)
  3. offers "Element", yeah that's sounds great, let's click "Continue"
  4. land on and then you need to register, aaargh
As you might have guessed by now, there is a specification to solve this, but web browsers need to adopt it as well, so that's far from actually being solved. At least browsers generally know about the matrix: scheme, it's just not exactly clear what they should do with it, especially when the handler is just another web page (e.g. Element web). In general, when compared with tools like Signal or WhatsApp, Matrix doesn't fare so well in terms of user discovery. I probably have some of my normal contacts that have a Matrix account as well, but there's really no way to know. It's kind of creepy when Signal tells you "this person is on Signal!" but it's also pretty cool that it works, and they actually implemented it pretty well. Registration is also less obvious: in Signal, the app confirms your phone number automatically. It's friction-less and quick. In Matrix, you need to learn about home servers, pick one, register (with a password! aargh!), and then setup encryption keys (not default), etc. It's a lot more friction. And look, I understand: giving away your phone number is a huge trade-off. I don't like it either. But it solves a real problem and makes encryption accessible to a ton more people. Matrix does have "identity servers" that can serve that purpose, but I don't feel confident sharing my phone number there. It doesn't help that the identity servers don't have private contact discovery: giving them your phone number is a more serious security compromise than with Signal. There's a catch-22 here too: because no one feels like giving away their phone numbers, no one does, and everyone assumes that stuff doesn't work anyways. Like it or not, Signal forcing people to divulge their phone number actually gives them critical mass that means actually a lot of my relatives are on Signal and I don't have to install crap like WhatsApp to talk with them.

5 minute clients evaluation Throughout all my tests I evaluated a handful of Matrix clients, mostly from Flathub because almost none of them are packaged in Debian. Right now I'm using Element, the flagship client from, in a web browser window, with the PopUp Window extension. This makes it look almost like a native app, and opens links in my main browser window (instead of a new tab in that separate window), which is nice. But I'm tired of buying memory to feed my web browser, so this indirection has to stop. Furthermore, I'm often getting completely logged off from Element, which means re-logging in, recovering my security keys, and reconfiguring my settings. That is extremely annoying. Coming from Irssi, Element is really "GUI-y" (pronounced "gooey"). Lots of clickety happening. To mark conversations as read, in particular, I need to click-click-click on all the tabs that have some activity. There's no "jump to latest message" or "mark all as read" functionality as far as I could tell. In Irssi the former is built-in (alt-a) and I made a custom /READ command for the latter:
/ALIAS READ script exec \$_->activity(0) for Irssi::windows
And yes, that's a Perl script in my IRC client. I am not aware of any Matrix client that does stuff like that, except maybe Weechat, if we can call it a Matrix client, or Irssi itself, now that it has a Matrix plugin (!). As for other clients, I have looked through the Matrix Client Matrix (confusing right?) to try to figure out which one to try, and, even after selecting Linux as a filter, the chart is just too wide to figure out anything. So I tried those, kind of randomly:
  • Fractal
  • Mirage
  • Nheko
  • Quaternion
Unfortunately, I lost my notes on those, I don't actually remember which one did what. I still have a session open with Mirage, so I guess that means it's the one I preferred, but I remember they were also all very GUI-y. Maybe I need to look at weechat-matrix or gomuks. At least Weechat is scriptable so I could continue playing the power-user. Right now my strategy with messaging (and that includes microblogging like Twitter or Mastodon) is that everything goes through my IRC client, so Weechat could actually fit well in there. Going with gomuks, on the other hand, would mean running it in parallel with Irssi or ... ditching IRC, which is a leap I'm not quite ready to take just yet. Oh, and basically none of those clients (except Nheko and Element) support VoIP, which is still kind of a second-class citizen in Matrix. It does not support large multimedia rooms, for example: Jitsi was used for FOSDEM instead of the native videoconferencing system.

Bots This falls a little aside the "usability" section, but I didn't know where to put this... There's a few Matrix bots out there, and you are likely going to be able to replace your existing bots with Matrix bots. It's true that IRC has a long and impressive history with lots of various bots doing various things, but given how young Matrix is, there's still a good variety:
  • maubot: generic bot with tons of usual plugins like sed, dice, karma, xkcd, echo, rss, reminder, translate, react, exec, gitlab/github webhook receivers, weather, etc
  • opsdroid: framework to implement "chat ops" in Matrix, connects with Matrix, GitHub, GitLab, Shell commands, Slack, etc
  • matrix-nio: another framework, used to build lots more bots like:
    • hemppa: generic bot with various functionality like weather, RSS feeds, calendars, cron jobs, OpenStreetmaps lookups, URL title snarfing, wolfram alpha, astronomy pic of the day, Mastodon bridge, room bridging, oh dear
    • devops: ping, curl, etc
    • podbot: play podcast episodes from AntennaPod
    • cody: Python, Ruby, Javascript REPL
    • eno: generic bot, "personal assistant"
  • mjolnir: moderation bot
  • hookshot: bridge with GitLab/GitHub
  • matrix-monitor-bot: latency monitor
One thing I haven't found an equivalent for is Debian's MeetBot. There's an archive bot but it doesn't have topics or a meeting chair, or HTML logs.

Working on Matrix As a developer, I find Matrix kind of intimidating. The specification is huge. The official specification itself looks somewhat digestable: it's only 6 APIs so that looks, at first, kind of reasonable. But whenever you start asking complicated questions about Matrix, you quickly fall into the Matrix Spec Change specification (which, yes, is a separate specification). And there are literally hundreds of MSCs flying around. It's hard to tell what's been adopted and what hasn't, and even harder to figure out if your specific client has implemented it. (One trendy answer to this problem is to "rewrite it in rust": Matrix are working on implementing a lot of those specifications in a matrix-rust-sdk that's designed to take the implementation details away from users.) Just taking the latest weekly Matrix report, you find that three new MSCs proposed, just last week! There's even a graph that shows the number of MSCs is progressing steadily, at 600+ proposals total, with the majority (300+) "new". I would guess the "merged" ones are at about 150. That's a lot of text which includes stuff like 3D worlds which, frankly, I don't think you should be working on when you have such important security and usability problems. (The internet as a whole, arguably, doesn't fare much better. RFC600 is a really obscure discussion about "INTERFACING AN ILLINOIS PLASMA TERMINAL TO THE ARPANET". Maybe that's how many MSCs will end up as well, left forgotten in the pits of history.) And that's the thing: maybe the Matrix people have a different objective than I have. They want to connect everything to everything, and make Matrix a generic transport for all sorts of applications, including virtual reality, collaborative editors, and so on. I just want secure, simple messaging. Possibly with good file transfers, and video calls. That it works with existing stuff is good, and it should be federated to remove the "Signal point of failure". So I'm a bit worried with the direction all those MSCs are taking, especially when you consider that clients other than Element are still struggling to keep up with basic features like end-to-end encryption or room discovery, never mind voice or spaces...

Conclusion Overall, Matrix is somehow in the space XMPP was a few years ago. It has a ton of features, pretty good clients, and a large community. It seems to have gained some of the momentum that XMPP has lost. It may have the most potential to replace Signal if something bad would happen to it (like, I don't know, getting banned or going nuts with cryptocurrency)... But it's really not there yet, and I don't see Matrix trying to get there either, which is a bit worrisome.

Looking back at history I'm also worried that we are repeating the errors of the past. The history of federated services is really fascinating:. IRC, FTP, HTTP, and SMTP were all created in the early days of the internet, and are all still around (except, arguably, FTP, which was removed from major browsers recently). All of them had to face serious challenges in growing their federation. IRC had numerous conflicts and forks, both at the technical level but also at the political level. The history of IRC is really something that anyone working on a federated system should study in detail, because they are bound to make the same mistakes if they are not familiar with it. The "short" version is:
  • 1988: Finnish researcher publishes first IRC source code
  • 1989: 40 servers worldwide, mostly universities
  • 1990: EFnet ("eris-free network") fork which blocks the "open relay", named Eris - followers of Eris form the A-net, which promptly dissolves itself, with only EFnet remaining
  • 1992: Undernet fork, which offered authentication ("services"), routing improvements and timestamp-based channel synchronisation
  • 1994: DALnet fork, from Undernet, again on a technical disagreement
  • 1995: Freenode founded
  • 1996: IRCnet forks from EFnet, following a flame war of historical proportion, splitting the network between Europe and the Americas
  • 1997: Quakenet founded
  • 1999: (XMPP founded)
  • 2001: 6 million users, OFTC founded
  • 2002: DALnet peaks at 136,000 users
  • 2003: IRC as a whole peaks at 10 million users, EFnet peaks at 141,000 users
  • 2004: (Facebook founded), Undernet peaks at 159,000 users
  • 2005: Quakenet peaks at 242,000 users, IRCnet peaks at 136,000 (Youtube founded)
  • 2006: (Twitter founded)
  • 2009: (WhatsApp, Pinterest founded)
  • 2010: (TextSecure AKA Signal, Instagram founded)
  • 2011: (Snapchat founded)
  • ~2013: Freenode peaks at ~100,000 users
  • 2016: IRCv3 standardisation effort started (TikTok founded)
  • 2021: Freenode self-destructs, Libera chat founded
  • 2022: Libera peaks at 50,000 users, OFTC peaks at 30,000 users
(The numbers were taken from the Wikipedia page and Note that I also include other networks launch in parenthesis for context.) Pretty dramatic, don't you think? Eventually, somehow, IRC became irrelevant for most people: few people are even aware of it now. With less than a million users active, it's smaller than Mastodon, XMPP, or Matrix at this point.1 If I were to venture a guess, I'd say that infighting, lack of a standardization body, and a somewhat annoying protocol meant the network could not grow. It's also possible that the decentralised yet centralised structure of IRC networks limited their reliability and growth. But large social media companies have also taken over the space: observe how IRC numbers peak around the time the wave of large social media companies emerge, especially Facebook (2.9B users!!) and Twitter (400M users).

Where the federated services are in history Right now, Matrix, and Mastodon (and email!) are at the "pre-EFnet" stage: anyone can join the federation. Mastodon has started working on a global block list of fascist servers which is interesting, but it's still an open federation. Right now, Matrix is totally open, but publishes a (federated) block list of hostile servers (, yes, of course it's a room). Interestingly, Email is also in that stage, where there are block lists of spammers, and it's a race between those blockers and spammers. Large email providers, obviously, are getting closer to the EFnet stage: you could consider they only accept email from themselves or between themselves. It's getting increasingly hard to deliver mail to Outlook and Gmail for example, partly because of bias against small providers, but also because they are including more and more machine-learning tools to sort through email and those systems are, fundamentally, unknowable. It's not quite the same as splitting the federation the way EFnet did, but the effect is similar. HTTP has somehow managed to live in a parallel universe, as it's technically still completely federated: anyone can start a web server if they have a public IP address and anyone can connect to it. The catch, of course, is how you find the darn thing. Which is how Google became one of the most powerful corporations on earth, and how they became the gatekeepers of human knowledge online. I have only briefly mentioned XMPP here, and my XMPP fans will undoubtedly comment on that, but I think it's somewhere in the middle of all of this. It was co-opted by Facebook and Google, and both corporations have abandoned it to its fate. I remember fondly the days where I could do instant messaging with my contacts who had a Gmail account. Those days are gone, and I don't talk to anyone over Jabber anymore, unfortunately. And this is a threat that Matrix still has to face. It's also the threat Email is currently facing. On the one hand corporations like Facebook want to completely destroy it and have mostly succeeded: many people just have an email account to register on things and talk to their friends over Instagram or (lately) TikTok (which, I know, is not Facebook, but they started that fire). On the other hand, you have corporations like Microsoft and Google who are still using and providing email services because, frankly, you still do need email for stuff, just like fax is still around but they are more and more isolated in their own silo. At this point, it's only a matter of time they reach critical mass and just decide that the risk of allowing external mail coming in is not worth the cost. They'll simply flip the switch and work on an allow-list principle. Then we'll have closed the loop and email will be dead, just like IRC is "dead" now. I wonder which path Matrix will take. Could it liberate us from these vicious cycles? Update: this generated some discussions on

  1. According to Wikipedia, there are currently about 500 distinct IRC networks operating, on about 1,000 servers, serving over 250,000 users. In contrast, Mastodon seems to be around 5 million users, claimed at FOSDEM 2021 to have about 28 million globally visible accounts, and Signal lays claim to over 40 million souls. XMPP claims to have "millions" of users on the homepage but the FAQ says they don't actually know. On the proprietary silo side of the fence, this page says
    • Facebook: 2.9 billion users
    • WhatsApp: 2B
    • Instagram: 1.4B
    • TikTok: 1B
    • Snapchat: 500M
    • Pinterest: 480M
    • Twitter: 397M
    Notable omission from that list: Youtube, with its mind-boggling 2.6 billion users... Those are not the kind of numbers you just "need to convince a brother or sister" to grow the network...

11 June 2022

Louis-Philippe V ronneau: Updating a rooted Pixel 3a

A short while after getting a Pixel 3a, I decided to root it, mostly to have more control over the charging procedure. In order to preserve battery life, I like my phone to stop charging at around 75% of full battery capacity and to shut down automatically at around 12%. Some Android ROMs have extra settings to manage this, but LineageOS unfortunately does not. Android already comes with a fairly complex mechanism to handle the charge cycle, but it is mostly controlled by the kernel and cannot be easily configured by end-users. acc is a higher-level "systemless" interface for the Android kernel battery management, but one needs root to do anything interesting with it. Once rooted, you can use the AccA app instead of playing on the command line to fine tune your battery settings. Sadly, having a rooted phone also means I need to re-root it each time there is an OS update (typically each week). Somehow, I keep forgetting the exact procedure to do this! Hopefully, I will be able to use this post as a reference in the future :) Note that these instructions might not apply to your exact phone model, proceed with caution! Extract the boot.img file This procedure mostly comes from the LineageOS documentation on extracting proprietary blobs from the payload.
  1. Download the latest LineageOS image for your phone.
  2. unzip the image to get the payload.bin file inside it.
  3. Clone the LineageOS scripts git repository: $ git clone
  4. extract the boot image (requires python3-protobuf): $ mkdir extracted-payload $ python3 scripts/update-payload-extractor/ payload.bin --output_dir extracted-payload
You should now have a boot.img file. Patch the boot image file using Magisk
  1. Upload the boot.img file you previously extracted to your device.
  2. Open Magisk and patch the boot.img file.
  3. Download the patched file back on your computer.
Flash the patched boot image
  1. Enable ADB debug mode on your phone.
  2. Reboot into fastboot mode. $ adb reboot fastboot
  3. Flash the patched boot image file: $ fastboot flash boot magisk_patched-foo.img
  4. Disable ADB debug mode on your phone.
Troubleshooting In an ideal world, you would do this entire process each time you upgrade to a new LineageOS version. Sadly, this creates friction and makes updating much more troublesome. To simplify things, you can try to flash an old patched boot.img file after upgrading, instead of generating it each time. In my experience, it usually works. When it does not, the device behaves weirdly after a reboot and things that require proprietary blobs (like WiFi) will stop working. If that happens:
  1. Download the latest LineageOS version for your phone.
  2. Reboot into recovery (Power + Volume Down).
  3. Click on "Apply Updates"
  4. Sideload the ROM: $ adb sideload

9 June 2022

Enrico Zini: Updating cbqt for bullseye

Back in 2017 I did work to setup a cross-building toolchain for QT Creator, that takes advantage of Debian's packaging for all the dependency ecosystem. It ended with cbqt which is a little script that sets up a chroot to hold cross-build-dependencies, to avoid conflicting with packages in the host system, and sets up a qmake alternative to make use of them. Today I'm dusting off that work, to ensure it works on Debian bullseye. Resetting QT Creator To make things reproducible, I wanted to reset QT Creator's configuration. Besides purging and reinstalling the package, one needs to manually remove:
  • ~/.config/QtProject
  • ~/.cache/QtProject/
  • /usr/share/qtcreator/QtProject which is where configuration is stored if you used sdktool to programmatically configure Qt Creator (see for example this post and see Debian bug #1012561.
Updating cbqt Easy start, change the distribution for the chroot:
-DIST_CODENAME = "stretch"
+DIST_CODENAME = "bullseye"
Adding LIBDIR Something else does not work:
Test$ qmake-armhf -makefile
Info: creating stash file  /Test/.qmake.stash
Test$ make
/usr/bin/arm-linux-gnueabihf-g++ -Wl,-O1 -Wl,-rpath-link, /armhf/lib/arm-linux-gnueabihf -Wl,-rpath-link, /armhf/usr/lib/arm-linux-gnueabihf -Wl,-rpath-link, /armhf/usr/lib/ -o Test main.o mainwindow.o moc_mainwindow.o    /armhf/usr/lib/arm-linux-gnueabihf/  /armhf/usr/lib/arm-linux-gnueabihf/  /armhf/usr/lib/arm-linux-gnueabihf/ -lGLESv2 -lpthread
/usr/lib/gcc-cross/arm-linux-gnueabihf/10/../../../../arm-linux-gnueabihf/bin/ld: cannot find -lGLESv2
collect2: error: ld returned 1 exit status
make: *** [Makefile:146: Test] Error 1
I figured that now I also need to set QMAKE_LIBDIR and not just QMAKE_RPATHLINKDIR:
--- a/cbqt
+++ b/cbqt
@@ -241,18 +241,21 @@ include(../common/linux.conf)
+QMAKE_LIBDIR +=  chroot.abspath /lib/arm-linux-gnueabihf
+QMAKE_LIBDIR +=  chroot.abspath /usr/lib/arm-linux-gnueabihf
+QMAKE_LIBDIR +=  chroot.abspath /usr/lib/
 QMAKE_RPATHLINKDIR +=  chroot.abspath /lib/arm-linux-gnueabihf
 QMAKE_RPATHLINKDIR +=  chroot.abspath /usr/lib/arm-linux-gnueabihf
 QMAKE_RPATHLINKDIR +=  chroot.abspath /usr/lib/
Now it links again:
Test$ qmake-armhf -makefile
Test$ make
/usr/bin/arm-linux-gnueabihf-g++ -Wl,-O1 -Wl,-rpath-link, /armhf/lib/arm-linux-gnueabihf -Wl,-rpath-link, /armhf/usr/lib/arm-linux-gnueabihf -Wl,-rpath-link, /armhf/usr/lib/ -o Test main.o mainwindow.o moc_mainwindow.o   -L /armhf/lib/arm-linux-gnueabihf -L /armhf/usr/lib/arm-linux-gnueabihf -L /armhf/usr/lib/  /armhf/usr/lib/arm-linux-gnueabihf/  /armhf/usr/lib/arm-linux-gnueabihf/  /armhf/usr/lib/arm-linux-gnueabihf/ -lGLESv2 -lpthread
Making it work in Qt Creator Time to try it in Qt Creator, and sadly it fails:
 /armhf/usr/lib/arm-linux-gnueabihf/qt5/mkspecs/features/toolchain.prf:76: Variable QMAKE_CXX.COMPILER_MACROS is not defined.
QMAKE_CXX.COMPILER_MACROS is not defined I traced it to this bit in armhf/usr/lib/arm-linux-gnueabihf/qt5/mkspecs/features/toolchain.prf (nonrelevant bits deleted):
isEmpty($$ target_prefix .COMPILER_MACROS)  
      else: gcc ghs  
        vars = $$qtVariablesFromGCC($$QMAKE_CXX)
    for (v, vars)  
        $$ target_prefix .COMPILER_MACROS += $$v
    cache($$ target_prefix .COMPILER_MACROS, set stash)
It turns out that qmake is not able to realise that the compiler is gcc, so vars does not get set, nothing is set in COMPILER_MACROS, and qmake fails. Reproducing it on the command line When run manually, however, qmake-armhf worked, so it would be good to know how Qt Creator is actually running qmake. Since it frustratingly does not show what commands it runs, I'll have to strace it:
strace -e trace=execve --string-limit=123456 -o qtcreator.trace -f qtcreator
And there it is:
$ grep qmake- qtcreator.trace
1015841 execve("/usr/local/bin/qmake-armhf", ["/usr/local/bin/qmake-armhf", "-query"], 0x56096e923040 /* 54 vars */) = 0
1015865 execve("/usr/local/bin/qmake-armhf", ["/usr/local/bin/qmake-armhf", " /Test/", "-spec", "arm-linux-gnueabihf", "CONFIG+=debug", "CONFIG+=qml_debug"], 0x7f5cb4023e20 /* 55 vars */) = 0
I run the command manually and indeed I reproduce the problem:
$ /usr/local/bin/qmake-armhf -spec arm-linux-gnueabihf CONFIG+=debug CONFIG+=qml_debug
 /armhf/usr/lib/arm-linux-gnueabihf/qt5/mkspecs/features/toolchain.prf:76: Variable QMAKE_CXX.COMPILER_MACROS is not defined.
I try removing options until I find the one that breaks it and... now it's always broken! Even manually running qmake-armhf, like I did earlier, stopped working:
$ rm .qmake.stash
$ qmake-armhf -makefile
 /armhf/usr/lib/arm-linux-gnueabihf/qt5/mkspecs/features/toolchain.prf:76: Variable QMAKE_CXX.COMPILER_MACROS is not defined.
Debugging toolchain.prf I tried purging and reinstalling qtcreator, and recreating the chroot, but qmake-armhf is staying broken. I'll let that be, and try to debug toolchain.prf. By grepping gcc in the mkspecs directory, I managed to figure out that:
  • The else: gcc ghs test is matching the value(s) of QMAKE_COMPILER
  • QMAKE_COMPILER can have multiple values, separated by space
  • If in armhf/usr/lib/arm-linux-gnueabihf/qt5/mkspecs/arm-linux-gnueabihf/qmake.conf I set QMAKE_COMPILER = gcc arm-linux-gnueabihf-gcc, then things work again.
Sadly, I failed to find reference documentation for QMAKE_COMPILER's syntax and behaviour. I also failed to find why qmake-armhf worked earlier, and I am also failing to restore the system to a situation where it works again. Maybe I dreamt that it worked? I had some manual change laying around from some previous fiddling with things? Anyway at least now I have the fix:
--- a/cbqt
+++ b/cbqt
@@ -248,7 +248,7 @@ QMAKE_RPATHLINKDIR +=  chroot.abspath /lib/arm-linux-gnueabihf
 QMAKE_RPATHLINKDIR +=  chroot.abspath /usr/lib/arm-linux-gnueabihf
 QMAKE_RPATHLINKDIR +=  chroot.abspath /usr/lib/
-QMAKE_COMPILER          =  chroot.arch_triplet -gcc
+QMAKE_COMPILER          = gcc  chroot.arch_triplet -gcc
 QMAKE_CC                = /usr/bin/ chroot.arch_triplet -gcc
Fixing a compiler mismatch warning In setting up the kit, Qt Creator also complained that the compiler from qmake did not match the one configured in the kit. That was easy to fix, by pointing at the host system cross-compiler in qmake.conf:
 QMAKE_COMPILER          =  chroot.arch_triplet -gcc
-QMAKE_CC                =  chroot.arch_triplet -gcc
+QMAKE_CC                = /usr/bin/ chroot.arch_triplet -gcc
 QMAKE_LINK_C            = $$QMAKE_CC
-QMAKE_CXX               =  chroot.arch_triplet -g++
+QMAKE_CXX               = /usr/bin/ chroot.arch_triplet -g++
 QMAKE_LINK              = $$QMAKE_CXX
Updated setup instructions Create an armhf environment:
sudo cbqt ./armhf --create --verbose
Create a qmake wrapper that builds with this environment:
sudo ./cbqt ./armhf --qmake -o /usr/local/bin/qmake-armhf
Install the build-dependencies that you need:
# Note: :arch is added automatically to package names if no arch is explicitly specified
sudo ./cbqt ./armhf --install libqt5svg5-dev libmosquittopp-dev qtwebengine5-dev
Build with qmake Use qmake-armhf instead of qmake and it works perfectly:
qmake-armhf -makefile
Set up Qt Creator Configure a new Kit in Qt Creator:
  1. Tools/Options, then Kits, then Add
  2. Name: armhf (or anything you like)
  3. In the Qt Versions tab, click Add then set the path of the new Qt to /usr/local/bin/qmake-armhf. Click Apply.
  4. Back in the Kits, select the Qt version you just created in the Qt version field
  5. In Compilers, select the ARM versions of GCC. If they do not appear, install crossbuild-essential-armhf, then in the Compilers tab click Re-detect and then Apply to make them available for selection
  6. Dismiss the dialog with "OK": the new kit is ready
Now you can choose the default kit to build and run locally, and the armhf kit for remote cross-development. I tried looking at sdktool to automate this step, and it requires a nontrivial amount of work to do it reliably, so these manual instructions will have to do. Credits This has been done as part of my work with Truelite.

8 June 2022

Laura Arjona Reina: Moving to a faster but smaller disk, encrypted setup

My work computer runs Debian 11 bullseye (the current stable release) in a mechanical 500GB disk, and I was provided with a new SDD disk but its size was 480 GB. So I had to shrink my partitions before copying the data to the new disk. It turned out to be a bit difficult because my main partition was encrypted. I write here how I did, maybe there are other simpler ways but I couldn t find them. References: I had three partitions in my old 500GB disk: /dev/sda1 is the EFI partition, /dev/sda2 the boot partition and /dev/sda3 the root partition (encrypted, with LVM, the standard way the Debian installer proposes when you choose a simple encrypted setup). First of all, I made a disk image with Clonezilla to an external USB disk, just in case I mess up things, to be able to return to a safe point and start again. Then I started my computer with a Debian 11 live USB with KDE Plasma desktop and Spanish localisation environment. I opened the KDE Partition manager and copied the non encrypted partitions (sda1, EFI and sda2, /boot) to the new disk. I shrinked the encrypted partition from the terminal with the following commands (I had enough free space so reduced my partition to a total of 300GB): Removed the swap partition and re-created it:
sudo lvremove /dev/larjona-pc-vg/swap_1
sudo pvresize --setphysicalvolumesize 380G /dev/mapper/cryptdisk
sudo pvchange -x y /dev/mapper/cryptdisk
sudo lvcreate -L 4G -n swap_1 larjona-pc-vg
sudo mkswap -L swap_1 /dev/larjona-pc-vg/swap_1
Display information about the physical volume in order to shrink it:
sudo pvs -v --segments --units s /dev/mapper/cryptdisk
sudo cryptsetup -b 838860800 resize cryptdisk
sudo cryptsetup status cryptdisk
sudo vgchange -a n vgroup
sudo vgchange -an
sudo cryptsetup luksClose cryptdisk
Then reduced the sda3 partition with the KDE partition manager (it took a while), and copy it to the new disk. Turned off the computer and unplugged the old disk. Started the computer with the Debian 11 Live USB again, UEFI boot. Now, to make my system boot:
sudo cryptsetup luksOpen /dev/sda3 crypdisk
sudo vgscan --mknodes
sudo vgchange -ay
sudo mount /dev/mapper/larjona--pc--vg-root /mnt
sudo mount /dev/sda2 /mnt/boot
sudo mount /dev/sda1 /mnt/boot/efi
mount --rbind /sys /media/linux/sys
mount -t efivarfs none /sys/firmware/efi/efivars
for i in /dev /dev/pts /proc /run; do sudo mount -B $i /mnt$i; done
sudo chroot /mnt
Then edited /mnt/etc/crypttab to reflect the name of the new encrypted partition, edited /mnt/etc/fstab to paste the UUIDs of the new partitions.
Then ran grub-install and reinstalled the kernels as noted in the reference, rebooted and logged in my Plasma desktop  (Well, the actual process was not so smooth but after several tries and errors and searching for help I managed to get the needed commands to make my system boot from the new disk).

26 May 2022

Sergio Talens-Oliag: New Blog Config

As promised, on this post I m going to explain how I ve configured this blog using hugo, asciidoctor and the papermod theme, how I publish it using nginx, how I ve integrated the remark42 comment system and how I ve automated its publication using gitea and json2file-go. It is a long post, but I hope that at least parts of it can be interesting for some, feel free to ignore it if that is not your case

Hugo Configuration

Theme settingsThe site is using the PaperMod theme and as I m using asciidoctor to publish my content I ve adjusted the settings to improve how things are shown with it. The current config.yml file is the one shown below (probably some of the settings are not required nor being used right now, but I m including the current file, so this post will have always the latest version of it):
title: Mixinet BlogOps
paginate: 5
theme: PaperMod
destination: public/
enableInlineShortcodes: true
enableRobotsTXT: true
buildDrafts: false
buildFuture: false
buildExpired: false
enableEmoji: true
pygmentsUseClasses: true
  disableXML: true
  minifyOutput: true
    languageName: "English"
    description: "Mixinet BlogOps -"
    author: "Sergio Talens-Oliag"
    weight: 1
    title: Mixinet BlogOps
      Title: "Sergio Talens-Oliag Technical Blog"
      Content: >
        ![Mixinet BlogOps](/images/mixinet-blogops.png)
      category: categories
      tag: tags
      series: series
        - name: Archive
          url: archives
          weight: 5
        - name: Categories
          url: categories/
          weight: 10
        - name: Tags
          url: tags/
          weight: 10
        - name: Search
          url: search/
          weight: 15
    - HTML
    - RSS
    - JSON
  env: production
  defaultTheme: light
  disableThemeToggle: false
  ShowShareButtons: true
  ShowReadingTime: true
  disableSpecial1stPost: true
  disableHLJS: true
  displayFullLangName: true
  ShowPostNavLinks: true
  ShowBreadCrumbs: true
  ShowCodeCopyButtons: true
  ShowRssButtonInSectionTermList: true
  ShowFullTextinRSS: true
  ShowToc: true
  TocOpen: false
  comments: true
  remark42SiteID: "blogops"
  remark42Url: "/remark42"
    enabled: false
    title: Sergio Talens-Oliag Technical Blog
    imageUrl: "/images/mixinet-blogops.png"
    imageTitle: Mixinet BlogOps
      - name: Archives
        url: archives
      - name: Categories
        url: categories
      - name: Tags
        url: tags
    - name: CV
      url: ""
    - name: Debian
      url: ""
    - name: GitHub
      url: ""
    - name: GitLab
      url: ""
    - name: Linkedin
      url: ""
    - name: RSS
      url: "index.xml"
    disableHLJS: true
    favicon: "/favicon.ico"
    favicon16x16:  "/favicon-16x16.png"
    favicon32x32:  "/favicon-32x32.png"
    apple_touch_icon:  "/apple-touch-icon.png"
    safari_pinned_tab:  "/safari-pinned-tab.svg"
    isCaseSensitive: false
    shouldSort: true
    location: 0
    distance: 1000
    threshold: 0.4
    minMatchCharLength: 0
    keys: ["title", "permalink", "summary", "content"]
    backend: html5s
    extensions: ['asciidoctor-html5s','asciidoctor-diagram']
    failureLevel: fatal
    noHeaderOrFooter: true
    preserveTOC: false
    safeMode: unsafe
    sectionNumbers: false
    trace: false
    verbose: false
    workingFolderCurrent: true
    disabled: false
    simple: true
    disabled: false
    enableDNT: true
    simple: true
    disabled: false
    simple: true
    disabled: false
    privacyEnhanced: true
    disableInlineCSS: true
    disableInlineCSS: true
      - '^asciidoctor$'
      - '^dart-sass-embedded$'
      - '^go$'
      - '^npx$'
      - '^postcss$'
Some notes about the settings:
  • disableHLJS and assets.disableHLJS are set to true; we plan to use rouge on adoc and the inclusion of the hljs assets adds styles that collide with the ones used by rouge.
  • ShowToc is set to true and the TocOpen setting is set to false to make the ToC appear collapsed initially. My plan was to use the asciidoctor ToC, but after trying I believe that the theme one looks nice and I don t need to adjust styles, although it has some issues with the html5s processor (the admonition titles use <h6> and they are shown on the ToC, which is weird), to fix it I ve copied the layouts/partial/toc.html to my site repository and replaced the range of headings to end at 5 instead of 6 (in fact 5 still seems a lot, but as I don t think I ll use that heading level on the posts it doesn t really matter).
  • params.profileMode values are adjusted, but for now I ve left it disabled setting params.profileMode.enabled to false and I ve set the homeInfoParams to show more or less the same content with the latest posts under it (I ve added some styles to my custom.css style sheet to center the text and image of the first post to match the look and feel of the profile).
  • On the asciidocExt section I ve adjusted the backend to use html5s, I ve added the asciidoctor-html5s and asciidoctor-diagram extensions to asciidoctor and adjusted the workingFolderCurrent to true to make asciidoctor-diagram work right (haven t tested it yet).

Theme customisationsTo write in asciidoctor using the html5s processor I ve added some files to the assets/css/extended directory:
  1. As said before, I ve added the file assets/css/extended/custom.css to make the homeInfoParams look like the profile page and I ve also changed a little bit some theme styles to make things look better with the html5s output:
    /* Fix first entry alignment to make it look like the profile */
    .first-entry   text-align: center;  
    .first-entry img   display: inline;  
     * Remove margin for .post-content code and reduce padding to make it look
     * better with the asciidoctor html5s output.
    .post-content code   margin: auto 0; padding: 4px;  
  2. I ve also added the file assets/css/extended/adoc.css with some styles taken from the asciidoctor-default.css, see this blog post about the original file; mine is the same after formatting it with css-beautify and editing it to use variables for the colors to support light and dark themes:
    /* AsciiDoctor*/
        border-collapse: collapse;
        border-spacing: 0
        border-collapse: separate;
        border: 0;
        background: none;
        width: 100%
    .admonitionblock>table td.icon  
        text-align: center;
        width: 80px
    .admonitionblock>table td.icon img  
        max-width: none
    .admonitionblock>table td.icon .title  
        font-weight: bold;
        font-family: "Open Sans", "DejaVu Sans", sans-serif;
        text-transform: uppercase
    .admonitionblock>table td.content  
        padding-left: 1.125em;
        padding-right: 1.25em;
        border-left: 1px solid #ddddd8;
        color: var(--primary)
    .admonitionblock>table td.content>:last-child>:last-child  
        margin-bottom: 0
    .admonitionblock td.icon [class^="fa icon-"]  
        font-size: 2.5em;
        text-shadow: 1px 1px 2px var(--secondary);
        cursor: default
    .admonitionblock td.icon .icon-note::before  
        content: "\f05a";
        color: var(--icon-note-color)
    .admonitionblock td.icon .icon-tip::before  
        content: "\f0eb";
        color: var(--icon-tip-color)
    .admonitionblock td.icon .icon-warning::before  
        content: "\f071";
        color: var(--icon-warning-color)
    .admonitionblock td.icon .icon-caution::before  
        content: "\f06d";
        color: var(--icon-caution-color)
    .admonitionblock td.icon .icon-important::before  
        content: "\f06a";
        color: var(--icon-important-color)
        display: inline-block;
        color: #fff !important;
        background-color: rgba(100, 100, 0, .8);
        -webkit-border-radius: 100px;
        border-radius: 100px;
        text-align: center;
        font-size: .75em;
        width: 1.67em;
        height: 1.67em;
        line-height: 1.67em;
        font-family: "Open Sans", "DejaVu Sans", sans-serif;
        font-style: normal;
        font-weight: bold
    .conum[data-value] *  
        color: #fff !important
        display: none
        content: attr(data-value)
    pre .conum[data-value]  
        position: relative;
        top: -.125em
    b.conum *  
        color: inherit !important
        display: none
  3. The previous file uses variables from a partial copy of the theme-vars.css file that changes the highlighted code background color and adds the color definitions used by the admonitions:
        /* Solarized base2 */
        /* --hljs-bg: rgb(238, 232, 213); */
        /* Solarized base3 */
        /* --hljs-bg: rgb(253, 246, 227); */
        /* Solarized base02 */
        --hljs-bg: rgb(7, 54, 66);
        /* Solarized base03 */
        /* --hljs-bg: rgb(0, 43, 54); */
        /* Default asciidoctor theme colors */
        --icon-note-color: #19407c;
        --icon-tip-color: var(--primary);
        --icon-warning-color: #bf6900;
        --icon-caution-color: #bf3400;
        --icon-important-color: #bf0000
        --hljs-bg: rgb(7, 54, 66);
        /* Asciidoctor theme colors with tint for dark background */
        --icon-note-color: #3e7bd7;
        --icon-tip-color: var(--primary);
        --icon-warning-color: #ff8d03;
        --icon-caution-color: #ff7847;
        --icon-important-color: #ff3030
  4. The previous styles use font-awesome, so I ve downloaded its resources for version 4.7.0 (the one used by asciidoctor) storing the font-awesome.css into on the assets/css/extended dir (that way it is merged with the rest of .css files) and copying the fonts to the static/assets/fonts/ dir (will be served directly):
    curl "$FA_BASE_URL/css/font-awesome.css" \
      > assets/css/extended/font-awesome.css
    for f in FontAwesome.otf fontawesome-webfont.eot \
      fontawesome-webfont.svg fontawesome-webfont.ttf \
      fontawesome-webfont.woff fontawesome-webfont.woff2; do
        curl "$FA_BASE_URL/fonts/$f" > "static/assets/fonts/$f"
  5. As already said the default highlighter is disabled (it provided a css compatible with rouge) so we need a css to do the highlight styling; as rouge provides a way to export them, I ve created the assets/css/extended/rouge.css file with the thankful_eyes theme:
    rougify style thankful_eyes > assets/css/extended/rouge.css
  6. To support the use of the html5s backend with admonitions I ve added a variation of the example found on this blog post to assets/js/adoc-admonitions.js:
    // replace the default admonitions block with a table that uses a format
    // similar to the standard asciidoctor ... as we are using fa-icons here there
    // is no need to add the icons: font entry on the document.
    window.addEventListener('load', function ()  
      const admonitions = document.getElementsByClassName('admonition-block')
      for (let i = admonitions.length - 1; i >= 0; i--)  
        const elm = admonitions[i]
        const type = elm.classList[1]
        const title = elm.getElementsByClassName('block-title')[0];
    	const label = title.getElementsByClassName('title-label')[0]
    		.innerHTML.slice(0, -1);
        const text = elm.innerHTML
        const parent = elm.parentNode
        const tempDiv = document.createElement('div')
        tempDiv.innerHTML =  <div class="admonitionblock $ type ">
              <td class="icon">
                <i class="fa icon-$ type " title="$ label "></i>
              <td class="content">
                $ text 
        const input = tempDiv.childNodes[0]
        parent.replaceChild(input, elm)
    and enabled its minified use on the layouts/partials/extend_footer.html file adding the following lines to it:
     - $admonitions := slice (resources.Get "js/adoc-admonitions.js")
        resources.Concat "assets/js/adoc-admonitions.js"   minify   fingerprint  
    <script defer crossorigin="anonymous" src="  $admonitions.RelPermalink  "
      integrity="  $admonitions.Data.Integrity  "></script>

Remark42 configurationTo integrate Remark42 with the PaperMod theme I ve created the file layouts/partials/comments.html with the following content based on the remark42 documentation, including extra code to sync the dark/light setting with the one set on the site:
<div id="remark42"></div>
  var remark_config =  
    host:   .Site.Params.remark42Url  ,
    site_id:   .Site.Params.remark42SiteID  ,
    url:   .Permalink  ,
    locale:   .Site.Language.Lang  
    /* Adjust the theme using the local-storage pref-theme if set */
    if (localStorage.getItem("pref-theme") === "dark")  
      remark_config.theme = "dark";
      else if (localStorage.getItem("pref-theme") === "light")  
      remark_config.theme = "light";
    /* Add remark42 widget */
    for(var i = 0; i < c.length; i++) 
      var d = document, s = d.createElement('script');
      s.src = + '/web/' + c[i] +'.js';
      s.defer = true;
      (d.head   d.body).appendChild(s);
   )(remark_config.components   ['embed']);
In development I use it with anonymous comments enabled, but to avoid SPAM the production site uses social logins (for now I ve only enabled Github & Google, if someone requests additional services I ll check them, but those were the easy ones for me initially). To support theme switching with remark42 I ve also added the following inside the layouts/partials/extend_footer.html file:
 - if (not site.Params.disableThemeToggle)  
/* Function to change theme when the toggle button is pressed */
document.getElementById("theme-toggle").addEventListener("click", () =>  
  if (typeof window.REMARK42 != "undefined")  
    if (document.body.className.includes('dark'))  
 - end  
With this code if the theme-toggle button is pressed we change the remark42 theme before the PaperMod one (that s needed here only, on page loads the remark42 theme is synced with the main one using the code from the layouts/partials/comments.html shown earlier).

Development setupTo preview the site on my laptop I m using docker-compose with the following configuration:
version: "2"
      context: ./docker/hugo-adoc
      dockerfile: ./Dockerfile
    image: sto/hugo-adoc
    container_name: hugo-adoc-blogops
    restart: always
      - .:/documents
    command: server --bind -D -F
    user: $ APP_UID :$ APP_GID 
    image: nginx:latest
    container_name: nginx-blogops
    restart: always
      - ./nginx/default.conf:/etc/nginx/conf.d/default.conf
      -  1313:1313
      context: ./docker/remark42
      dockerfile: ./Dockerfile
    image: sto/remark42
    container_name: remark42-blogops
    restart: always
      - ./.env
      - ./remark42/
      - ./remark42/
To run it properly we have to create the .env file with the current user ID and GID on the variables APP_UID and APP_GID (if we don t do it the files can end up being owned by a user that is not the same as the one running the services):
$ echo "APP_UID=$(id -u)\nAPP_GID=$(id -g)" > .env
The Dockerfile used to generate the sto/hugo-adoc is:
FROM asciidoctor/docker-asciidoctor:latest
RUN gem install --no-document asciidoctor-html5s &&\
 apk update && apk add --no-cache curl libc6-compat &&\
 repo_path="gohugoio/hugo" &&\
 api_url="$repo_path/releases/latest" &&\
  curl -sL "$api_url"  \
  sed -n "s/^.*download_url\": \"\\(.*.extended.*Linux-64bit.tar.gz\)\"/\1/p"\
 )" &&\
 curl -sL "$download_url" -o /tmp/hugo.tgz &&\
 tar xf /tmp/hugo.tgz hugo &&\
 install hugo /usr/bin/ &&\
 rm -f hugo /tmp/hugo.tgz &&\
 /usr/bin/hugo version &&\
 apk del curl && rm -rf /var/cache/apk/*
# Expose port for live server
ENTRYPOINT ["/usr/bin/hugo"]
CMD [""]
If you review it you will see that I m using the docker-asciidoctor image as the base; the idea is that this image has all I need to work with asciidoctor and to use hugo I only need to download the binary from their latest release at github (as we are using an image based on alpine we also need to install the libc6-compat package, but once that is done things are working fine for me so far). The image does not launch the server by default because I don t want it to; in fact I use the same docker-compose.yml file to publish the site in production simply calling the container without the arguments passed on the docker-compose.yml file (see later). When running the containers with docker-compose up (or docker compose up if you have the docker-compose-plugin package installed) we also launch a nginx container and the remark42 service so we can test everything together. The Dockerfile for the remark42 image is the original one with an updated version of the script:
FROM umputun/remark42:latest
The updated is similar to the original, but allows us to use an APP_GID variable and updates the /etc/group file of the container so the files get the right user and group (with the original script the group is always 1001):
#!/sbin/dinit /bin/sh
uid="$(id -u)"
if [ "$ uid " -eq "0" ]; then
  echo "init container"
  # set container's time zone
  cp "/usr/share/zoneinfo/$ TIME_ZONE " /etc/localtime
  echo "$ TIME_ZONE " >/etc/timezone
  echo "set timezone $ TIME_ZONE  ($(date))"
  # set UID & GID for the app
  if [ "$ APP_UID " ]   [ "$ APP_GID " ]; then
    [ "$ APP_UID " ]   APP_UID="1001"
    [ "$ APP_GID " ]   APP_GID="$ APP_UID "
    echo "set custom APP_UID=$ APP_UID  & APP_GID=$ APP_GID "
    sed -i "s/^app:x:1001:1001:/app:x:$ APP_UID :$ APP_GID :/" /etc/passwd
    sed -i "s/^app:x:1001:/app:x:$ APP_GID :/" /etc/group
    echo "custom APP_UID and/or APP_GID not defined, using 1001:1001"
  chown -R app:app /srv /home/app
echo "prepare environment"
# replace  % REMARK_URL %  by content of REMARK_URL variable
find /srv -regex '.*\.\(html\ js\ mjs\)$' -print \
  -exec sed -i "s % REMARK_URL % $ REMARK_URL  g"   \;
if [ -n "$ SITE_ID " ]; then
  #replace "site_id: 'remark'" by SITE_ID
  sed -i "s 'remark' '$ SITE_ID ' g" /srv/web/*.html
echo "execute \"$*\""
if [ "$ uid " -eq "0" ]; then
  exec su-exec app "$@"
  exec "$@"
The environment file used with remark42 for development is quite minimal:
And the nginx/default.conf file used to publish the service locally is simple too:
 listen 1313;
 server_name localhost;
 location /  
    proxy_pass http://hugo:1313;
    proxy_set_header Host $http_host;
    proxy_set_header X-Real-IP $remote_addr;
    proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
    proxy_set_header X-Forwarded-Proto $scheme;
    proxy_set_header Upgrade $http_upgrade;
    proxy_set_header Connection "upgrade";
 location /remark42/  
    rewrite /remark42/(.*) /$1 break;
    proxy_pass http://remark42:8080/;
    proxy_set_header Host $http_host;
    proxy_set_header X-Real-IP $remote_addr;
    proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
    proxy_set_header X-Forwarded-Proto $scheme;

Production setupThe VM where I m publishing the blog runs Debian GNU/Linux and uses binaries from local packages and applications packaged inside containers. To run the containers I m using docker-ce (I could have used podman instead, but I already had it installed on the machine, so I stayed with it). The binaries used on this project are included on the following packages from the main Debian repository:
  • git to clone & pull the repository,
  • jq to parse json files from shell scripts,
  • json2file-go to save the webhook messages to files,
  • inotify-tools to detect when new files are stored by json2file-go and launch scripts to process them,
  • nginx to publish the site using HTTPS and work as proxy for json2file-go and remark42 (I run it using a container),
  • task-spool to queue the scripts that update the deployment.
And I m using docker and docker compose from the debian packages on the docker repository:
  • docker-ce to run the containers,
  • docker-compose-plugin to run docker compose (it is a plugin, so no - in the name).

Repository checkoutTo manage the git repository I ve created a deploy key, added it to gitea and cloned the project on the /srv/blogops PATH (that route is owned by a regular user that has permissions to run docker, as I said before).

Compiling the site with hugoTo compile the site we are using the docker-compose.yml file seen before, to be able to run it first we build the container images and once we have them we launch hugo using docker compose run:
$ cd /srv/blogops
$ git pull
$ docker compose build
$ if [ -d "./public" ]; then rm -rf ./public; fi
$ docker compose run hugo --
The compilation leaves the static HTML on /srv/blogops/public (we remove the directory first because hugo does not clean the destination folder as jekyll does). The deploy script re-generates the site as described and moves the public directory to its final place for publishing.

Running remark42 with dockerOn the /srv/blogops/remark42 folder I have the following docker-compose.yml:
version: "2"
      context: ../docker/remark42
      dockerfile: ./Dockerfile
    image: sto/remark42
      - ../.env
      - ./
    container_name: remark42
    restart: always
      - ./
The ../.env file is loaded to get the APP_UID and APP_GID variables that are used by my version of the script to adjust file permissions and the file contains the rest of the settings for remark42, including the social network tokens (see the remark42 documentation for the available parameters, I don t include my configuration here because some of them are secrets).

Nginx configurationThe nginx configuration for the site is as simple as:
  listen 443 ssl http2;
  ssl_certificate /etc/letsencrypt/live/;
  ssl_certificate_key /etc/letsencrypt/live/;
  include /etc/letsencrypt/options-ssl-nginx.conf;
  ssl_dhparam /etc/letsencrypt/ssl-dhparams.pem;
  access_log /var/log/nginx/;
  error_log  /var/log/nginx/;
  root /srv/blogops/nginx/public_html;
  location /  
    try_files $uri $uri/ =404;
  include /srv/blogops/nginx/remark42.conf;
  listen 80 ;
  listen [::]:80 ;
  access_log /var/log/nginx/;
  error_log  /var/log/nginx/;
  if ($host =  
    return 301 https://$host$request_uri;
  return 404;
On this configuration the certificates are managed by certbot and the server root directory is on /srv/blogops/nginx/public_html and not on /srv/blogops/public; the reason for that is that I want to be able to compile without affecting the running site, the deployment script generates the site on /srv/blogops/public and if all works well we rename folders to do the switch, making the change feel almost atomic.

json2file-go configurationAs I have a working WireGuard VPN between the machine running gitea at my home and the VM where the blog is served, I m going to configure the json2file-go to listen for connections on a high port using a self signed certificate and listening on IP addresses only reachable through the VPN. To do it we create a systemd socket to run json2file-go and adjust its configuration to listen on a private IP (we use the FreeBind option on its definition to be able to launch the service even when the IP is not available, that is, when the VPN is down). The following script can be used to set up the json2file-go configuration:
set -e
# ---------
# ---------
# ----
# ----
# Install packages used with json2file for the blogops site
sudo apt update
sudo apt install -y json2file-go uuid
if [ -z "$(type mkcert)" ]; then
  sudo apt install -y mkcert
sudo apt clean
# Configuration file values
J2F_USER="$(id -u)"
J2F_GROUP="$(id -g)"
# Configure json2file
[ -d "$J2F_DIR" ]   mkdir "$J2F_DIR"
sudo sh -c "echo '$J2F_DIR' >'$J2F_BASEDIR_FILE'"
[ -d "$TLS_DIR" ]   mkdir "$TLS_DIR"
if [ ! -f "$J2F_CRT_PATH" ]   [ ! -f "$J2F_KEY_PATH" ]; then
  mkcert -cert-file "$J2F_CRT_PATH" -key-file "$J2F_KEY_PATH" "$(hostname -f)"
sudo sh -c "echo '$J2F_CRT_PATH' >'$J2F_CRT_FILE'"
sudo sh -c "echo '$J2F_KEY_PATH' >'$J2F_KEY_FILE'"
sudo sh -c "cat >'$J2F_DIRLIST_FILE'" <<EOF
$(echo "$J2F_DIRLIST"   tr ';' '\n')
# Service override
[ -d "$J2F_SERVICE_DIR" ]   sudo mkdir "$J2F_SERVICE_DIR"
sudo sh -c "cat >'$J2F_SERVICE_OVERRIDE'" <<EOF
# Socket override
[ -d "$J2F_SOCKET_DIR" ]   sudo mkdir "$J2F_SOCKET_DIR"
sudo sh -c "cat >'$J2F_SOCKET_OVERRIDE'" <<EOF
# Set FreeBind to listen on missing addresses (the VPN can be down sometimes)
# Set ListenStream to nothing to clear its value and add the new value later
# Restart and enable service
sudo systemctl daemon-reload
sudo systemctl stop "$J2F_SERVICE_NAME"
sudo systemctl start "$J2F_SERVICE_NAME"
sudo systemctl enable "$J2F_SERVICE_NAME"
# ----
# vim: ts=2:sw=2:et:ai:sts=2
Warning: The script uses mkcert to create the temporary certificates, to install the package on bullseye the backports repository must be available.

Gitea configurationTo make gitea use our json2file-go server we go to the project and enter into the hooks/gitea/new page, once there we create a new webhook of type gitea and set the target URL to and on the secret field we put the token generated with uuid by the setup script:
sed -n -e 's/blogops://p' /etc/json2file-go/dirlist
The rest of the settings can be left as they are:
  • Trigger on: Push events
  • Branch filter: *
Warning: We are using an internal IP and a self signed certificate, that means that we have to review that the webhook section of the app.ini of our gitea server allows us to call the IP and skips the TLS verification (you can see the available options on the gitea documentation). The [webhook] section of my server looks like this:
Once we have the webhook configured we can try it and if it works our json2file server will store the file on the /srv/blogops/webhook/json2file/blogops/ folder.

The json2file spooler scriptWith the previous configuration our system is ready to receive webhook calls from gitea and store the messages on files, but we have to do something to process those files once they are saved in our machine. An option could be to use a cronjob to look for new files, but we can do better on Linux using inotify we will use the inotifywait command from inotify-tools to watch the json2file output directory and execute a script each time a new file is moved inside it or closed after writing (IN_CLOSE_WRITE and IN_MOVED_TO events). To avoid concurrency problems we are going to use task-spooler to launch the scripts that process the webhooks using a queue of length 1, so they are executed one by one in a FIFO queue. The spooler script is this:
set -e
# ---------
# ---------
# ---------
# ---------
  echo "Queuing job to process file '$1'"
    tsp -n "$WEBHOOK_COMMAND" "$1"
# ----
# ----
if [ ! -d "$INPUT_DIR" ]; then
  echo "Input directory '$INPUT_DIR' does not exist, aborting!"
  exit 1
[ -d "$TSP_DIR" ]   mkdir "$TSP_DIR"
echo "Processing existing files under '$INPUT_DIR'"
find "$INPUT_DIR" -type f   sort   while read -r _filename; do
  queue_job "$_filename"
# Use inotifywatch to process new files
echo "Watching for new files under '$INPUT_DIR'"
inotifywait -q -m -e close_write,moved_to --format "%w%f" -r "$INPUT_DIR"  
  while read -r _filename; do
    queue_job "$_filename"
# ----
# vim: ts=2:sw=2:et:ai:sts=2
To run it as a daemon we install it as a systemd service using the following script:
set -e
# ---------
# ---------
# Configuration file values
J2F_USER="$(id -u)"
J2F_GROUP="$(id -g)"
# ----
# ----
# Install packages used with the webhook processor
sudo apt update
sudo apt install -y inotify-tools jq task-spooler
sudo apt clean
# Configure process service
sudo sh -c "cat > $SPOOLER_SERVICE_FILE" <<EOF
Description=json2file processor for $J2F_USER
# Restart and enable service
sudo systemctl daemon-reload
sudo systemctl stop "$SPOOLER_SERVICE_NAME"   true
sudo systemctl start "$SPOOLER_SERVICE_NAME"
sudo systemctl enable "$SPOOLER_SERVICE_NAME"
# ----
# vim: ts=2:sw=2:et:ai:sts=2

The gitea webhook processorFinally, the script that processes the JSON files does the following:
  1. First, it checks if the repository and branch are right,
  2. Then, it fetches and checks out the commit referenced on the JSON file,
  3. Once the files are updated, compiles the site using hugo with docker compose,
  4. If the compilation succeeds the script renames directories to swap the old version of the site by the new one.
If there is a failure the script aborts but before doing it or if the swap succeeded the system sends an email to the configured address and/or the user that pushed updates to the repository with a log of what happened. The current script is this one:
set -e
# ---------
# ---------
# Values
# Address that gets all messages, leave it empty if not wanted
# If the following variable is set to 'true' the pusher gets mail on failures
# If the following variable is set to 'true' the pusher gets mail on success
# gitea's conf/app.ini value of NO_REPLY_ADDRESS, it is used for email domains
# when the KeepEmailPrivate option is enabled for a user
# Directories
# Files
TODAY="$(date +%Y%m%d)"
OUTPUT_BASENAME="$(date +%Y%m%d-%H%M%S.%N)"
# Query to get variables from a gitea webhook json
  printf "%s" \
    '(.             @sh "gt_ref=\(.ref);"),' \
    '(.             @sh "gt_after=\(.after);"),' \
    '(.repository   @sh "gt_repo_clone_url=\(.clone_url);"),' \
    '(.repository   @sh "gt_repo_name=\(.name);"),' \
    '(.pusher       @sh "gt_pusher_full_name=\(.full_name);"),' \
    '(.pusher       @sh "gt_pusher_email=\(.email);")'
# ---------
# Functions
# ---------
  echo "$(date -R) $*" >>"$WEBHOOK_LOGFILE_PATH"
    [ -d "$_d" ]   mkdir "$_d"
  # Try to remove empty dirs
    if [ -d "$_d" ]; then
      rmdir "$_d" 2>/dev/null   true
  webhook_log "Accepted: $*"
  webhook_log "Rejected: $*"
  if [ -f "$WEBHOOK_JSON_INPUT_FILE" ]; then
  exit 0
  webhook_log "Deployed: $*"
  webhook_log "Troubled: $*"
  # Add the pusher email address unless it is from the domain NO_REPLY_ADDRESS,
  # which should match the value of that variable on the gitea 'app.ini' (it
  # is the domain used for emails when the user hides it).
  # shellcheck disable=SC2154
  if [ -n "$ gt_pusher_email##*@"$ NO_REPLY_ADDRESS " " ] &&
    [ -z "$ gt_pusher_email##*@* " ]; then
    _user_email="\"$gt_pusher_full_name <$gt_pusher_email>\""
  if [ "$_addr" ] && [ "$_user_email" ]; then
    echo "$_addr,$_user_email"
  elif [ "$_user_email" ]; then
    echo "$_user_email"
  elif [ "$_addr" ]; then
    echo "$_addr"
  if [ "$MAIL_LOGFILE" = "true" ]; then
    to_addr="$(print_mailto "$to_addr")"
  if [ "$to_addr" ]; then
    # shellcheck disable=SC2154
    subject="OK - $gt_repo_name updated to commit '$gt_after'"
    mail -s "$ MAIL_PREFIX $ subject " "$to_addr" \
  if [ "$MAIL_ERRFILE" = true ]; then
    to_addr="$(print_mailto "$to_addr")"
  if [ "$to_addr" ]; then
    # shellcheck disable=SC2154
    subject="KO - $gt_repo_name update FAILED for commit '$gt_after'"
    mail -s "$ MAIL_PREFIX $ subject " "$to_addr" \
# ----
# ----
# Check directories
# Go to the base directory
cd "$BASE_DIR"
# Check if the file exists
if [ ! -f "$WEBHOOK_JSON_INPUT_FILE" ]; then
  webhook_reject "Input arg '$1' is not a file, aborting"
# Parse the file
webhook_log "Processing file '$WEBHOOK_JSON_INPUT_FILE'"
# Check that the repository clone url is right
# shellcheck disable=SC2154
if [ "$gt_repo_clone_url" != "$REPO_CLONE_URL" ]; then
  webhook_reject "Wrong repository: '$gt_clone_url'"
# Check that the branch is the right one
# shellcheck disable=SC2154
if [ "$gt_ref" != "$REPO_REF" ]; then
  webhook_reject "Wrong repository ref: '$gt_ref'"
# Accept the file
# shellcheck disable=SC2154
webhook_accept "Processing '$gt_repo_name'"
# Update the checkout
git fetch >>"$WEBHOOK_LOGFILE_PATH" 2>&1   ret="$?"
if [ "$ret" -ne "0" ]; then
  webhook_troubled "Repository fetch failed"
# shellcheck disable=SC2154
git checkout "$gt_after" >>"$WEBHOOK_LOGFILE_PATH" 2>&1   ret="$?"
if [ "$ret" -ne "0" ]; then
  webhook_troubled "Repository checkout failed"
# Remove the build dir if present
if [ -d "$PUBLIC_DIR" ]; then
  rm -rf "$PUBLIC_DIR"
# Build site
docker compose run hugo -- >>"$WEBHOOK_LOGFILE_PATH" 2>&1   ret="$?"
# go back to the main branch
git switch main && git pull
# Fail if public dir was missing
if [ "$ret" -ne "0" ]   [ ! -d "$PUBLIC_DIR" ]; then
  webhook_troubled "Site build failed"
# Remove old public_html copies
webhook_log 'Removing old site versions, if present'
find $NGINX_BASE_DIR -mindepth 1 -maxdepth 1 -name 'public_html-*' -type d \
  -exec rm -rf   \; >>"$WEBHOOK_LOGFILE_PATH" 2>&1   ret="$?"
if [ "$ret" -ne "0" ]; then
  webhook_troubled "Removal of old site versions failed"
# Switch site directory
TS="$(date +%Y%m%d-%H%M%S)"
if [ -d "$PUBLIC_HTML_DIR" ]; then
  webhook_log "Moving '$PUBLIC_HTML_DIR' to '$PUBLIC_HTML_DIR-$TS'"
if [ "$ret" -eq "0" ]; then
  webhook_log "Moving '$PUBLIC_DIR' to '$PUBLIC_HTML_DIR'"
if [ "$ret" -ne "0" ]; then
  webhook_troubled "Site switch failed"
  webhook_deployed "Site deployed successfully"
# ----
# vim: ts=2:sw=2:et:ai:sts=2

27 April 2022

Antoine Beaupr : building Debian packages under qemu with sbuild

I've been using sbuild for a while to build my Debian packages, mainly because it's what is used by the Debian autobuilders, but also because it's pretty powerful and efficient. Configuring it just right, however, can be a challenge. In my quick Debian development guide, I had a few pointers on how to configure sbuild with the normal schroot setup, but today I finished a qemu based configuration.

Why I want to use qemu mainly because it provides better isolation than a chroot. I sponsor packages sometimes and while I typically audit the source code before building, it still feels like the extra protection shouldn't hurt. I also like the idea of unifying my existing virtual machine setup with my build setup. My current VM is kind of all over the place: libvirt, vagrant, GNOME Boxes, etc?). I've been slowly converging over libvirt however, and most solutions I use right now rely on qemu under the hood, certainly not chroots... I could also have decided to go with containers like LXC, LXD, Docker (with conbuilder, whalebuilder, docker-buildpackage), systemd-nspawn (with debspawn), unshare (with schroot --chroot-mode=unshare), or whatever: I didn't feel those offer the level of isolation that is provided by qemu. The main downside of this approach is that it is (obviously) slower than native builds. But on modern hardware, that cost should be minimal.

How Basically, you need this:
sudo mkdir -p /srv/sbuild/qemu/
sudo apt install sbuild-qemu
sudo sbuild-qemu-create -o /srv/sbuild/qemu/unstable.img unstable
Then to make this used by default, add this to ~/.sbuildrc:
# run autopkgtest inside the schroot
$run_autopkgtest = 1;
# tell sbuild to use autopkgtest as a chroot
$chroot_mode = 'autopkgtest';
# tell autopkgtest to use qemu
$autopkgtest_virt_server = 'qemu';
# tell autopkgtest-virt-qemu the path to the image
# use --debug there to show what autopkgtest is doing
$autopkgtest_virt_server_options = [ '--', '/srv/sbuild/qemu/%r-%a.img' ];
# tell plain autopkgtest to use qemu, and the right image
$autopkgtest_opts = [ '--', 'qemu', '/srv/sbuild/qemu/%r-%a.img' ];
# no need to cleanup the chroot after build, we run in a completely clean VM
$purge_build_deps = 'never';
# no need for sudo
$autopkgtest_root_args = '';
Note that the above will use the default autopkgtest (1GB, one core) and qemu (128MB, one core) configuration, which might be a little low on resources. You probably want to be explicit about this, with something like this:
# extra parameters to pass to qemu
# --enable-kvm is not necessary, detected on the fly by autopkgtest
my @_qemu_options = ['--ram-size=4096', '--cpus=2'];
# tell autopkgtest-virt-qemu the path to the image
# use --debug there to show what autopkgtest is doing
$autopkgtest_virt_server_options = [ @_qemu_options, '--', '/srv/sbuild/qemu/%r-%a.img' ];
$autopkgtest_opts = [ '--', 'qemu', @qemu_options, '/srv/sbuild/qemu/%r-%a.img'];
This configuration will:
  1. create a virtual machine image in /srv/sbuild/qemu for unstable
  2. tell sbuild to use that image to create a temporary VM to build the packages
  3. tell sbuild to run autopkgtest (which should really be default)
  4. tell autopkgtest to use qemu for builds and for tests
Note that the VM created by sbuild-qemu-create have an unlocked root account with an empty password.

Other useful tasks
  • enter the VM to make test, changes will be discarded (thanks Nick Brown for the sbuild-qemu-boot tip!):
     sbuild-qemu-boot /srv/sbuild/qemu/unstable-amd64.img
    That program is shipped only with bookworm and later, an equivalent command is:
     qemu-system-x86_64 -snapshot -enable-kvm -object rng-random,filename=/dev/urandom,id=rng0 -device virtio-rng-pci,rng=rng0,id=rng-device0 -m 2048 -nographic /srv/sbuild/qemu/unstable-amd64.img
    The key argument here is -snapshot.
  • enter the VM to make permanent changes, which will not be discarded:
     sudo sbuild-qemu-boot --readwrite /srv/sbuild/qemu/unstable-amd64.img
    Equivalent command:
     sudo qemu-system-x86_64 -enable-kvm -object rng-random,filename=/dev/urandom,id=rng0 -device virtio-rng-pci,rng=rng0,id=rng-device0 -m 2048 -nographic /srv/sbuild/qemu/unstable-amd64.img
  • update the VM (thanks lavamind):
     sudo sbuild-qemu-update /srv/sbuild/qemu/unstable-amd64.img
  • build in a specific VM regardless of the suite specified in the changelog (e.g. UNRELEASED, bookworm-backports, bookworm-security, etc):
     sbuild --autopkgtest-virt-server-opts="-- qemu /var/lib/sbuild/qemu/bookworm-amd64.img"
    Note that you'd also need to pass --autopkgtest-opts if you want autopkgtest to run in the correct VM as well:
     sbuild --autopkgtest-opts="-- qemu /var/lib/sbuild/qemu/unstable.img" --autopkgtest-virt-server-opts="-- qemu /var/lib/sbuild/qemu/bookworm-amd64.img"
    You might also need parameters like --ram-size if you customized it above.
And yes, this is all quite complicated and could be streamlined a little, but that's what you get when you have years of legacy and just want to get stuff done. It seems to me autopkgtest-virt-qemu should have a magic flag starts a shell for you, but it doesn't look like that's a thing. When that program starts, it just says ok and sits there. Maybe because the authors consider the above to be simple enough (see also bug #911977 for a discussion of this problem).

Live access to a running test When autopkgtest starts a VM, it uses this funky qemu commandline:
qemu-system-x86_64 -m 4096 -smp 2 -nographic -net nic,model=virtio -net user,hostfwd=tcp: -object rng-random,filename=/dev/urandom,id=rng0 -device virtio-rng-pci,rng=rng0,id=rng-device0 -monitor unix:/tmp/autopkgtest-qemu.w1mlh54b/monitor,server,nowait -serial unix:/tmp/autopkgtest-qemu.w1mlh54b/ttyS0,server,nowait -serial unix:/tmp/autopkgtest-qemu.w1mlh54b/ttyS1,server,nowait -virtfs local,id=autopkgtest,path=/tmp/autopkgtest-qemu.w1mlh54b/shared,security_model=none,mount_tag=autopkgtest -drive index=0,file=/tmp/autopkgtest-qemu.w1mlh54b/overlay.img,cache=unsafe,if=virtio,discard=unmap,format=qcow2 -enable-kvm -cpu kvm64,+vmx,+lahf_lm
... which is a typical qemu commandline, I'm sorry to say. That gives us a VM with those settings (paths are relative to a temporary directory, /tmp/autopkgtest-qemu.w1mlh54b/ in the above example):
  • the shared/ directory is, well, shared with the VM
  • port 10022 is forward to the VM's port 22, presumably for SSH, but not SSH server is started by default
  • the ttyS1 and ttyS2 UNIX sockets are mapped to the first two serial ports (use nc -U to talk with those)
  • the monitor UNIX socket is a qemu control socket (see the QEMU monitor documentation, also nc -U)
In other words, it's possible to access the VM with:
nc -U /tmp/autopkgtest-qemu.w1mlh54b/ttyS2
The nc socket interface is ... not great, but it works well enough. And you can probably fire up an SSHd to get a better shell if you feel like it.

Nitty-gritty details no one cares about

Fixing hang in sbuild cleanup I'm having a hard time making heads or tails of this, but please bear with me. In sbuild + schroot, there's this notion that we don't really need to cleanup after ourselves inside the schroot, as the schroot will just be delted anyways. This behavior seems to be handled by the internal "Session Purged" parameter. At least in lib/Sbuild/, we can see this:
my $is_cloned_session = (defined ($session->get('Session Purged')) &&
             $session->get('Session Purged') == 1) ? 1 : 0;
if ($is_cloned_session)  
$self->log("Not cleaning session: cloned chroot in use\n");
if ($purge_build_deps)  
    # Removing dependencies
    $self->log("Not removing build depends: as requested\n");
The schroot builder defines that parameter as:
    $self->set('Session Purged', $info-> 'Session Purged' );
... which is ... a little confusing to me. $info is:
my $info = $self->get('Chroots')->get_info($schroot_session);
... so I presume that depends on whether the schroot was correctly cleaned up? I stopped digging there... is way more explicit:
$self->set('Session Purged', 1);
I wonder if we should do something like this with the autopkgtest backend. I guess people might technically use it with something else than qemu, but qemu is the typical use case of the autopkgtest backend, in my experience. Or at least certainly with things that cleanup after themselves. Right? For some reason, before I added this line to my configuration:
$purge_build_deps = 'never';
... the "Cleanup" step would just completely hang. It was quite bizarre.

Disgression on the diversity of VM-like things There are a lot of different virtualization solutions one can use (e.g. Xen, KVM, Docker or Virtualbox). I have also found libguestfs to be useful to operate on virtual images in various ways. Libvirt and Vagrant are also useful wrappers on top of the above systems. There are particularly a lot of different tools which use Docker, Virtual machines or some sort of isolation stronger than chroot to build packages. Here are some of the alternatives I am aware of: Take, for example, Whalebuilder, which uses Docker to build packages instead of pbuilder or sbuild. Docker provides more isolation than a simple chroot: in whalebuilder, packages are built without network access and inside a virtualized environment. Keep in mind there are limitations to Docker's security and that pbuilder and sbuild do build under a different user which will limit the security issues with building untrusted packages. On the upside, some of things are being fixed: whalebuilder is now an official Debian package (whalebuilder) and has added the feature of passing custom arguments to dpkg-buildpackage. None of those solutions (except the autopkgtest/qemu backend) are implemented as a sbuild plugin, which would greatly reduce their complexity. I was previously using Qemu directly to run virtual machines, and had to create VMs by hand with various tools. This didn't work so well so I switched to using Vagrant as a de-facto standard to build development environment machines, but I'm returning to Qemu because it uses a similar backend as KVM and can be used to host longer-running virtual machines through libvirt. The great thing now is that autopkgtest has good support for qemu and sbuild has bridged the gap and can use it as a build backend. I originally had found those bugs in that setup, but all of them are now fixed:
  • #911977: sbuild: how do we correctly guess the VM name in autopkgtest?
  • #911979: sbuild: fails on chown in autopkgtest-qemu backend
  • #911963: autopkgtest qemu build fails with proxy_cmd: parameter not set
  • #911981: autopkgtest: qemu server warns about missing CPU features
So we have unification! It's possible to run your virtual machines and Debian builds using a single VM image backend storage, which is no small feat, in my humble opinion. See the sbuild-qemu blog post for the annoucement Now I just need to figure out how to merge Vagrant, GNOME Boxes, and libvirt together, which should be a matter of placing images in the right place... right? See also hosting.

pbuilder vs sbuild I was previously using pbuilder and switched in 2017 to sbuild. has a good comparative between pbuilder and sbuild that shows they are pretty similar. The big advantage of sbuild is that it is the tool in use on the buildds and it's written in Perl instead of shell. My concerns about switching were POLA (I'm used to pbuilder), the fact that pbuilder runs as a separate user (works with sbuild as well now, if the _apt user is present), and setting up COW semantics in sbuild (can't just plug cowbuilder there, need to configure overlayfs or aufs, which was non-trivial in Debian jessie). Ubuntu folks, again, have more documentation there. Debian also has extensive documentation, especially about how to configure overlays. I was ultimately convinced by stapelberg's post on the topic which shows how much simpler sbuild really is...

Who Thanks lavamind for the introduction to the sbuild-qemu package.

8 April 2022

Jacob Adams: The Unexpected Importance of the Trailing Slash

For many using Unix-derived systems today, we take for granted that /some/path and /some/path/ are the same. Most shells will even add a trailing slash for you when you press the Tab key after the name of a directory or a symbolic link to one. However, many programs treat these two paths as subtly different in certain cases, which I outline below, as all three have tripped me up in various ways1.

POSIX and Coreutils Perhaps the trickiest use of the trailing slash in a distinguishing way is in POSIX2 which states:
When the final component of a pathname is a symbolic link, the standard requires that a trailing <slash> causes the link to be followed. This is the behavior of historical implementations3. For example, for /a/b and /a/b/, if /a/b is a symbolic link to a directory, then /a/b refers to the symbolic link, and /a/b/ refers to the directory to which the symbolic link points.
This leads to some unexpected behavior. For example, if you have the following structure of a directory dir containing a file dirfile with a symbolic link link pointing to dir. (which will be used in all shell examples throughout this article):
$ ls -lR
total 4
drwxr-xr-x 2 jacob jacob 4096 Apr  3 00:00 dir
lrwxrwxrwx 1 jacob jacob    3 Apr  3 00:00 link -> dir
total 0
-rw-r--r-- 1 jacob jacob 0 Apr  3 00:12 dirfile
On Unixes such as MacOS, FreeBSD or Illumos4, you can move a directory through a symbolic link by using a trailing slash:
$ mv link/ otherdir
$ ls
link	otherdir
On Linux5, mv will not rename the indirectly referenced directory and not the symbolic link, when given a symbolic link with a trailing slash as the source to be renamed. despite the coreutils documentation s claims to the contrary6, instead failing with Not a directory:
$ mv link/ other
mv: cannot move 'link/' to 'other': Not a directory
$ mkdir otherdir
$ mv link/ otherdir
mv: cannot move 'link/' to 'otherdir/link': Not a directory
$ mv link/ otherdir/
mv: cannot move 'link/' to 'otherdir/link': Not a directory
$ mv link otherdirlink
$ ls -l otherdirlink
lrwxrwxrwx 1 jacob jacob 3 Apr  3 00:13 otherdirlink -> dir
This is probably for the best, as it is very confusing behavior. There is still one advantage the trailing slash has when using mv, even on Linux, in that is it does not allow you to move a file to a non-existent directory, or move a file that you expect to be a directory that isn t.
$ mv dir/dirfile nonedir/
mv: cannot move 'dir/dirfile' to 'nonedir/': Not a directory
$ touch otherfile
$ mv otherfile/ dir
mv: cannot stat 'otherfile/': Not a directory
$ mv otherfile dir
$ ls dir
dirfile  otherfile
However, Linux still exhibits some confusing behavior of its own, like when you attempt to remove link recursively with a trailing slash:
rm -rvf link/
Neither link nor dir are removed, but the contents of dir are removed:
removed 'link/dirfile'
Whereas if you remove the trailing slash, you just remove the symbolic link:
$ rm -rvf link
removed 'link'
While on MacOS, FreeBSD or Illumos4, rm will also remove the source directory:
$ rm -rvf link
$ ls
The find and ls commands, in contrast, behave the same on all three operating systems. The find command only searches the contents of the directory a symbolic link points to if the trailing slash is added:
$ find link -name dirfile
$ find link/ -name dirfile
The ls command acts similarly, showing information on just a symbolic link by itself unless a trailing slash is added, at which point it shows the contents of the directory that it links to:
$ ls -l link
lrwxrwxrwx 1 jacob jacob 3 Apr  3 00:13 link -> dir
$ ls -l link/
total 0
-rw-r--r-- 1 jacob jacob 0 Apr  3 00:13 dirfile

rsync The command rsync handles a trailing slash in an unusual way that trips up many new users. The rsync man page notes:
You can think of a trailing / on a source as meaning copy the contents of this directory as opposed to copy the directory by name , but in both cases the attributes of the containing directory are transferred to the containing directory on the destination.
That is to say, if we had two folders a and b each of which contained some files:
$ ls -R .
a  b
a1  a2
b1  b2
Running rsync -av a b moves the entire directory a to directory b:
$ rsync -av a b
sending incremental file list
sent 181 bytes  received 58 bytes  478.00 bytes/sec
total size is 0  speedup is 0.00
$ ls -R b
a  b1  b2
a1  a2
While running rsync -av a/ b moves the contents of directory a to b:
$ rsync -av a/ b
sending incremental file list
sent 170 bytes  received 57 bytes  454.00 bytes/sec
total size is 0  speedup is 0.00
$ ls b
a1  a2	b1  b2

Dockerfile COPY The Dockerfile COPY command also cares about the presence of the trailing slash, using it to determine whether the destination should be considered a file or directory. The Docker documentation explains the rules of the command thusly:
COPY [--chown=<user>:<group>] <src>... <dest>
If <src> is a directory, the entire contents of the directory are copied, including filesystem metadata. Note: The directory itself is not copied, just its contents. If <src> is any other kind of file, it is copied individually along with its metadata. In this case, if <dest> ends with a trailing slash /, it will be considered a directory and the contents of <src> will be written at <dest>/base(<src>). If multiple <src> resources are specified, either directly or due to the use of a wildcard, then <dest> must be a directory, and it must end with a slash /. If <dest> does not end with a trailing slash, it will be considered a regular file and the contents of <src> will be written at <dest>. If <dest> doesn t exist, it is created along with all missing directories in its path.
This means if you had a COPY command that moved file to a nonexistent containerfile without the slash, it would create containerfile as a file with the contents of file.
COPY file /containerfile
container$ stat -c %F containerfile
regular empty file
Whereas if you add a trailing slash, then file will be added as a file under the new directory containerdir:
COPY file /containerdir/
container$ stat -c %F containerdir
Interestingly, at no point can you copy a directory completely, only its contents. Thus if you wanted to make a directory in the new container, you need to specify its name in both the source and the destination:
COPY dir /dirincontainer
container$ stat -c %F /dirincontainer
Dockerfiles do also make good use of the trailing slash to ensure they re doing what you mean by requiring a trailing slash on the destination of multiple files:
COPY file otherfile /othercontainerdir
results in the following error:
When using COPY with more than one source file, the destination must be a directory and end with a /
  1. I m sure there are probably more than just these three cases, but these are the three I m familiar with. If you know of more, please tell me about them!.
  2. Some additional relevant sections are the Path Resolution Appendix and the section on Symbolic Links.
  3. The sentence This is the behavior of historical implementations implies that this probably originated in some ancient Unix derivative, possibly BSD or even the original Unix. I don t really have a source on that though, so please reach out if you happen to have any more knowledge on what this refers to.
  4. I tested on MacOS 11.6.5, FreeBSD 12.0 and OmniOS 5.11 2
  5. unless the source is a directory trailing slashes give -ENOTDIR
  6. In fairness to the coreutils maintainers, it seems to be true on all other Unix platforms, but it probably deserves a mention in the documentation when Linux is the most common platform on which coreutils is used. I should submit a patch.

5 April 2022

Dirk Eddelbuettel: RcppArmadillo on CRAN: Upstream Updates

armadillo image Armadillo is a powerful and expressive C++ template library for linear algebra aiming towards a good balance between speed and ease of use with a syntax deliberately close to a Matlab. RcppArmadillo integrates this library with the R environment and language and is widely used by (currently) 972 other packages on CRAN, downloaded over 24 million times (per the partial logs from the cloud mirrors of CRAN), and the CSDA paper (preprint / vignette) by Conrad and myself has been cited 465 times according to Google Scholar. This release brings a new upstream release 11.0.0. We tested this very rigorously via three different RC release each of which got a full reverse-dependencies run (for which results are always logged here). The full set of changes (since the last CRAN release follows.

Changes in RcppArmadillo version (2022-04-04)
  • Upgraded to Armadillo release 11.0.0 (Creme Brulee)
    • added variants of inv() and inv_sympd() that provide rcond (reciprocal condition number)
    • expanded inv() and inv_sympd() with options inv_opts::tiny and inv_opts::allow_approx
    • stricter handling of singular matrices by inv() and inv_sympd()
    • stricter handling of non-sympd matrices by inv() and inv_sympd()
    • stricter handling of non-finitie matrices by pinv()
    • more robust handling of rank deficient matrices by solve()
    • faster handling of diagonal matrices by rcond()
    • changed eigs_sym() and eigs_gen() to use higher quality RNG
    • quantile() and median() will now throw an exception if given matrices/vectors have NaN elements
    • workaround for yet another bug in Intel MKL
  • Until May 2022, protect correction to Field behavior via define of RCPP_ARMADILLO_FIX_Field
  • If a LAPACK installation with missing complex routines is found (as e.g. Ubuntu using 3.9.0) then the LAPACK unit test is skipped.

Changes in RcppArmadillo version (2022-02-01)
  • Upgraded to Armadillo release 10.8.2 (Realm Raider)
    • fix potential speed regression in pinv() and rank()

Courtesy of my CRANberries, there is a diffstat report relative to previous release. More detailed information is on the RcppArmadillo page. Questions, comments etc should go to the rcpp-devel mailing list off the R-Forge page. If you like this or other open-source work I do, you can sponsor me at GitHub.

This post by Dirk Eddelbuettel originated on his Thinking inside the box blog. Please report excessive re-aggregation in third-party for-profit settings.

1 April 2022

Russell Coker: Converting to UEFI

When I got my HP ML110 Gen9 working as a workstation I initially was under the impression that boot wasn t supported on NVMe and booted it from USB. I found USB booting with legacy boot to be unreliable so decided to try EFI booting and noticed that the NVMe devices were boot candidates with UEFI. Making one of them bootable was more complex than expected because no-one seems to have documented such things. So here s my documentation, it s not great but this method has worked once for me. Before starting major partitioning work it s best to run parted -l and save the output to a file, that can allow you to recreate partitions if you corrupt them. One thing I m doing on systems I manage is putting @reboot /usr/sbin/parted -l > /root/parted.log in the root crontab, then when the system is backed up the backup server gets any recent changes to partitioning (I don t backup /var/log on all my systems). Firstly run parted on the device to create the EFI and /boot partitions, note that if you want to copy and paste from this you must do so one line at a time, a block paste seemed to confuse parted.
mklabel gpt
mkpart EFI fat32 1 99
mkpart boot ext3 99 300
toggle 1 boot
toggle 1 esp
# Model: CT1000P1SSD8 (nvme)
# Disk /dev/nvme1n1: 1000GB
# Sector size (logical/physical): 512B/512B
# Partition Table: gpt
# Disk Flags: 
# Number  Start   End     Size    File system  Name  Flags
#  1      1049kB  98.6MB  97.5MB  fat32        EFI   boot, esp
#  2      98.6MB  300MB   201MB   ext3         boot
Here are the commands needed to create the filesystems and install the necessary files. This is almost to the stage of being scriptable. Some minor changes need to be made to convert from NVMe device names to SATA/SAS but nothing serious.
mkfs.vfat /dev/nvme1n1p1
mkfs.ext3 -N 1000 /dev/nvme1n1p2
file -s /dev/nvme1n1p2   sed -e s/^.*UUID/UUID/ -e "s/ .*$/ \/boot ext3 noatime 0 1/" >> /etc/fstab
file -s /dev/nvme1n1p1   tr "[a-f]" "[A-F]"  sed -e s/^.*numBEr.0x/UUID=/ -e "s/, .*$/ \/boot\/efi vfat umask=0077 0 1/" >> /etc/fstab
# edit /etc/fstab to put a hyphen between the 2 groups of 4 chars for the VFAT filesystem UUID
mount /boot
mkdir -p /boot/efi /boot/grub
mount /boot/efi
mkdir -p /boot/efi/EFI/debian
apt install efibootmgr shim-unsigned grub-efi-amd64
cp /usr/lib/shim/* /usr/lib/grub/x86_64-efi/monolithic/grubx64.efi /boot/efi/EFI/debian
file -s /dev/nvme1n1p2   sed -e "s/^.*UUID=/search.fs_uuid /" -e "s/ .needs.*$/ root hd0,gpt2/" > /boot/efi/EFI/debian/grub.cfg
echo "set prefix=(\$root)'/boot/grub'" >> /boot/efi/EFI/debian/grub.cfg
echo "configfile \$prefix/grub.cfg" >> /boot/efi/EFI/debian/grub.cfg
If someone would like to make a script that can handle the different partition names of regular SCSI/SATA disks, NVMe, CCISS, etc then that would be great. It would be good to have a script in Debian that creates the partitions and sets up the EFI files. If you want to have a second bootable device then the following commands will copy a GPT partition table and give it new UUIDs, make very certain that $DISKB is the one you want to be wiped and refer to my previous mention of parted -l . Also note that parted has a rescue command which works very well.
sgdisk /dev/$DISKA -R /dev/$DISKB 
sgdisk -G /dev/$DISKB
To backup a GPT partition table run a command like this. Note that if sgdisk is told to backup a MBR partitioned disk it will say Found invalid GPT and valid MBR; converting MBR to GPT forma which is probably a viable way of converting MBR format to GPT.
sgdisk -b sda.bak /dev/sda