2024-11-27

Found three PCs in the Elektromüll (electronical trash) at HfK Bremen. I found a HP Pavilion ze4300, a Dell Optiplex 745 and a Hyrican PC PCK02282. You can read more about them in the [[ssw_device-list|device list]].

I tried to turn on the HP Pavilion and Dell Optiplex 745 but both of them did not start properly. Both of them are missing crucial components such as RAM, Hard Disks or Graphics Cards. The HP Pavilion is specifically missing RAM and a Hard Disk. The Dell Optiplex is also missing a Hard Disk and a Graphics Cards.

I found four 1GB Ram Sticks in the studio and added two of them to the Optiplex 745. This did not help with the startup process. Tomorrow I will read the manual and try to figure out the status codes.

2024-11-28

Added the two other 1GB Ram Sticks I found yesterday to the Hyrican PC. Then I tried to boot, startup seemed normal, got a screen signal and was routed into BIOS.

It shows that:

• NVRAM check passed

• USB controllers initialized

• 2040MB RAM detected

• 1 keyboard, 1 mouse, 4 storage devices connected

• CD-ROM drive (PHILIPS SPD6005T)

• 4 USB card readers (SD, CF, SM, MS formats)

• CMOS settings are incorrect

• Date/Time not set

Went into the SETUP and fixed the CMOS Settings, which was basically the time and date setting. After I fixed that I was asked to insert a Boot device.

In the meantime Lars said that he has a spare 4GB RAm Stick for the HP Pavilion laptop I found yesterday. He also said that his flatmate has a 8GB Ram Stick that he wants to gift away. He might send me both of them.

As no storage was found in the Hyrican PC I went to the DM Studio storage to look if I can find one in there. I found a 160GB HDD and connected it to the PC.

Went into the BIOS and booted from the Western Digital Harddrive. On there was a Windows 7 installation that started without any issues.

This allowed me to check the system settings to see what chip I have in the PC.

After that, I inserted a USB Flash Drive with Fedora 41 into the PC. I changed the Boot order in the BIOS and booted from the flash drive. I selected the option to test the environment and then install fedora. The PC got stuck in this process for a while only displaying a cursor.

After approx. 10min the tests started and almost failed immediately with the following log:

...

Checking: 004.8%

The media check is complete, the result is: FAIL

It is not recommended to use this media.
[FAILED] Failed to start checkisomd5@dev-sdf.service - Media check on /dev/sdf.
[ 155.341909] dracut-initqueue[1445]: Job for checkisomd5@dev-sdf.service failed because the control process exited with error code.
[ 155.342271] dracut-initqueue[1445]: See "systemctl status checkisomd5@dev-sdf.service" and "journalctl -xeu checkisomd5@dev-sdf.service" for details.
[ 155.394629] dracut-initqueue[1282]: Warning: Media check failed! We do not recommend using this medium. System will halt in 12 hours

Apparently this references some sort of error with the boot flash drive. I flashed the drive again.

I research the minimal requirements for Fedora 41 Server to see if the hardware requirements are sufficient. Even though a setup with 1GiB or 2GiB should work a setup with at least 4GiB is recommended.

Currently I have two 1GB Ram sticks inside the Hyrican PC. I will try it with them first and if this does not work I will upgrade to two 2GiB Ram sticks that I found in the DM Studio material storage.

The medium test for the Fedora boot drive took ages, therefore I skipped it. Then I did not have enough time to finish the installation. Will try again tomorrow.

2024-11-29

Read about Alpine Linux yesterday, which seems to be a very lightweight Linux distribution with minimal resource usage. Maybe worth a try.

Also Debian could be an option as it uses less resources then Fedora. For now I will test it with Fedora again.

While installing Fedora I did a bit more research about lightweight Linux distros and stumbled across AntiX Linux and TinyCore Linux, which also promote low resource usage. Especially AntiX Linux looks interesting because it is made to be booted from flash drives. Maybe something I should test after the Fedora install was successful.

The install of Fedora Server Linux was complete after roughly 1.5 hours without testing the flash drive first. Just direct install.

After the install I first upgraded all packages.

sudo dnf upgrade --refresh -y 

Then I installed NGINX following this guide. The connection to the server failed though.

SELinux is enabled which I could check with:

getenforce

Then I needed to run the following to allow http and https requests.

sudo setsebool -P httpd_can_network_connect 1

After that requesting the website inside the local network worked.

With Fedora Server an nginx running we have these stats:

top - 13:50:48 up  2:22,  3 users,  load average: 0,07, 0,04, 0,01
Tasks: 167 total,   1 running, 166 sleeping,   0 stopped,   0 zombie
%Cpu(s):  0,2 us,  0,2 sy,  0,0 ni, 99,3 id,  0,3 wa,  0,0 hi,  0,0 si,  0,0 st
MiB Mem :   1950,6 total,    610,5 free,    466,6 used,   1048,5 buff/cache
MiB Swap:   1950,0 total,   1945,7 free,      4,2 used.   1484,0 avail Mem

Together with Clemens I also found a PowerMac G4 MMD with (almost) max stats.

Seems like the power supply is broken though and apparently Apple used a proprietary 24PIN Layout for their power supplies. So let’s see if this can be fixed somehow.

2024-12-09

I ordered some parts from different people to complete the computers I found. The list includes:

• 2x 512MB SODIMM DDR Ram Sticks
  • For the HP Pavilion
• 1x 250GB SATA HDD from WD
  • For the Dell Optiplex
• 1x 160GB 2.5-Inch PATA Notebook Hard Drive
  • For the HP Pavilion
• 1x 80GB 2.5-Inch PATA Notebook Hard Drive
  • For the HP Pavillion

The two notebook hard drives were only sold as a package. Included to the two ram sticks was also a CPU chip, which unfortunately has bent socket pins.

2024-12-12

I installed Void Linux inside a UTM VM on my MacBook. Setting up was a bit tedious as I needed to setup everything from file system, boot partitions, etc.

Also when I tried to install lighttpd as a package I could not install it, because the distro/VM could not connect to the network.

The resource usage was around 200MB Ram which is already pretty good. I also got an SSD drive from my father today, which he does not need anymore.

2024-12-16

With all the hardware I bought I started to fix the laptop first. I installed the two 512MB RAM Sticks which allowed me to boot into BIOS.

After that I wanted to install one of the two IDE/PATA Hard Drives into the laptop. I did not fit immediately and I learned, that an adapter between the hard drive and the laptop. There is an image of it in the manual.

The adapter that is missing is part of the piece 319417-001 - Hard Drive Tray for 1F (One Fan), which can be for example be bought here and here but can not be shipped to Germany.

After a discussion with Clemens we came up with the idea to replace the optical drive with an optical drive to hard drive caddy adapter.

Something similar to this: To do so, I had to remove the optical drive first from the laptop. I looked into the manual on how to do that. I first removed the keyboard cover and then the keyboard itself. After that I could unscrew the optical drive and push it out.

After some research and googling I found out that the optical drive I have here is a 12.7mm one with an IDE/PATA interface. More specifically it is the 319422-001 – CD-RW/DVD Combo Assy.

All the hard drive caddy adapters I could find only had an SATA interface and not the needed IDE/PATA interface, which the optical drive had. Therefore I need to find another solution to add storage to this laptop.

One idea could be to open it up more to see if it has SATA ports or some IDE/PATA Ports I can add an adapter to. Maybe the floppy drive could work?

I could also buy this: https://www.ebay.com/itm/284524589152

2024-12-17

Installed an SSD drive into the Dell Optiplex 745 and now it stays on at least. I am not getting an image signal though.

I had two RAM sticks installed which where had the ECC (Error Correction). The RAM sticks were two of these models: Samsung M391T2863QZ3-CE66. From the [[manual-dell-optiplex-745.pdf|Dell Optiplex 745 manual]] I learned that only Non-ECC ram sticks are allowed (see page 145). Therefore I swapped the two RAM sticks with another one of this type: Samsung M378T5663RZ3-CF7. As 2GB is the maximum the motherboard supports I only installed one.

With the RAM stick swapped, I got a video signal and could enter the boot menu.

2024-12-18

After some thoughts and also chatting with various chat bots (😅) I decided to use Alpine as my distro of choice. Together with Ansible I should be able to create a fairly re-producible workflow for all three machines.

Setting up Alpine directly through Ansible does not really make sense. Using the setup script ansible-setup ist just way more convenient and faster.

2025-01-05

Found two spare notebooks at my parents place, that where supposed to go to the bin. The first one was an ASUS A9RP-5057H. Unfortunately after turning it nothing appeared on the screen. Tested it with the charged battery in it and without.

The second notebook was a Dell Inspiron 6400. That one booted on start slowly but flawlessly into Windows 8. As this notebook was still running out of the box I will move forward with this one for the thesis.

2025-01-07

I tried to install Fedora 42 on the Dell Inspiron 6400. Did not work as the Inspiron 6400 has a 32bit architecture while the Fedora ISO image I had was for 64bit systems.

Then I installed a Debian ISO Image for 32bit systems on the laptop.

2025-01-08

I thought a bit about what I still want to develop/test or do with the servers before starting to develop the website that they should serve.

It basically comes down to these three things:

2025-01-09

Wrote an Ansible playbook to deploy the hono/deno server for the nodes. Deployed it for now on a VM setup through UTM on my Mac.

2025-02-21

Clemens fixed the PowerPC Mac G4 by re-soldering the old heat of the power supply to a new power supply. Now the Mac works and we could even boot it correctly and open up MacOS.

2025-03-15

As I had problems with setting up audio on alpine, more precisely recording from a microphone, I researched a bit.

Consolidating some articles I should do the following steps:

• Install DBus and start a DBus session
  • https://wiki.alpinelinux.org/wiki/PipeWire
  • https://www.reddit.com/r/AlpineLinux/comments/plf465/how_to_setup_audio_on_alpine/
• Add user to audio and video group
• Add pipewire & wireplumber
• Maybe setup ALSA
  • Install pipewire-alsa
  • https://wiki.alpinelinux.org/wiki/ALSA
• Start the pipewire-launcher /usr/libexec/pipewire-launcher
• Add export $(dbus-launch) /usr/libexec/pipewire-launcher to the shell configuration

sudo apk add dbus

sudo rc-update add dbus

sudo rc-service dbus start

sudo apk add alsa-utils alsaconf

sudo addgroup $USER audio
sudo addgroup root audio

alsamixer

# Unmute and set Master volume to 80% 
sudo amixer -c 0 sset Master unmute 
sudo amixer -c 0 sset Master 80% 

# Unmute and set Speaker volume to 80% 
sudo amixer -c 0 sset Speaker 
unmute sudo amixer -c 0 sset Speaker 80% 

# Unmute and set Headphone volume to 80% 
sudo amixer -c 0 sset Headphone unmute 
sudo amixer -c 0 sset Headphone 80% 

# Unmute and set microphone (if needed) 
sudo amixer -c 0 sset Mic unmute 
sudo amixer -c 0 sset Mic 70% 
sudo amixer -c 0 sset "Mic Boost" 50% 

# Adjust Line if needed 
sudo amixer -c 0 sset Line unmute 
sudo amixer -c 0 sset Line 80% 

# Unmute and set PCM volume to 80% (if it exists) 
sudo amixer -c 0 sset PCM unmute 2>/dev/null 
sudo amixer -c 0 sset PCM 80% 2>/dev/null

# Store the settings 
sudo alsactl store

sudo vi /usr/share/alsa/alsa.conf

defaults.ctl.card 0
defaults.pcm.card 0

rc-service alsa start
rc-update add alsa

#### Ensuring microphones are always enabled for capture

# 1. Enable capture for all available inputs
sudo amixer -c 0 sset "Capture" cap
sudo amixer -c 0 sset "Capture" 100%
sudo amixer -c 0 sset "Mic Boost" 100%

# 2. Make sure Mic is selected as the capture source
sudo amixer -c 0 sset "Mic" cap
sudo amixer -c 0 sset "Line" cap  # Enable Line input as fallback

# 3. Create startup script to ensure mics are always enabled
cat > /etc/local.d/enable-mics.start << 'EOF'
#!/bin/sh

# Wait for sound system to initialize
sleep 2

# Enable all capture devices
amixer -c 0 sset "Capture" cap
amixer -c 0 sset "Capture" 100%
amixer -c 0 sset "Mic Boost" 100%
amixer -c 0 sset "Mic" cap
amixer -c 0 sset "Line" cap

# Create a test recording to verify mic is working
arecord -d 1 -f S16_LE -r 44100 -c 2 /tmp/mic-test.wav 2>/dev/null || true

# Log result
echo "Microphone initialization completed at $(date)" >> /var/log/mic-setup.log

exit 0
EOF

# 4. Make the startup script executable and add to startup
sudo chmod +x /etc/local.d/enable-mics.start
sudo rc-update add local default

# 5. Create an ALSA state file that ensures mics are enabled at boot
sudo alsactl store

# 6. Add custom rules to ALSA configuration to prioritize recording
cat > /etc/asound.conf << 'EOF'
pcm.!default {
    type asym
    playback.pcm "playback"
    capture.pcm "capture"
}

pcm.playback {
    type plug
    slave.pcm "hw:0,0"
}

pcm.capture {
    type plug
    slave.pcm "hw:0,0"
}

# Prioritize capture
defaults.pcm.!card 0
defaults.pcm.!device 0
defaults.ctl.!card 0
EOF

# 7. Reboot to ensure all changes take effect
echo "Configuration complete. Reboot system to apply all changes."

sudo apk add pipewire wireplumber pipewire-alsa

sudo mkdir -p /etc/pipewire/

sudo tee /etc/pipewire/pipewire.conf <<EOF
context.properties = {
    core.daemon = true
    core.name = pipewire-0
    default.clock.rate = 48000
    default.clock.quantum = 1024
    default.clock.min-quantum = 32
    default.clock.max-quantum = 8192
}

context.spa-libs = {
    audio.convert.* = audioconvert/libspa-audioconvert
    api.alsa.* = alsa/libspa-alsa
    api.v4l2.* = v4l2/libspa-v4l2
    api.libcamera.* = libcamera/libspa-libcamera
}

context.modules = [
    { name = libpipewire-module-rt
        args = {
            nice.level = -11
        }
        flags = [ ifexists nofail ]
    }
    { name = libpipewire-module-protocol-native }
    { name = libpipewire-module-profiler }
    { name = libpipewire-module-metadata }
    { name = libpipewire-module-spa-device-factory }
    { name = libpipewire-module-spa-node-factory }
    { name = libpipewire-module-client-node }
    { name = libpipewire-module-client-device }
    { name = libpipewire-module-access
        args = {
            # Allow access to all users in the "audio" group
            access.allowed = [ "*" ]
            access.rejected = [ ]
            access.allowed-groups = [ "audio" ]
        }
    }
    { name = libpipewire-module-adapter }
    { name = libpipewire-module-link-factory }
    { name = libpipewire-module-session-manager }
    { name = libpipewire-module-portal
        flags = [ ifexists nofail ]
    }
]
EOF

sudo mkdir -p /etc/wireplumber/main.lua.d/

sudo tee /etc/wireplumber/main.lua.d/51-alsa-config.lua <<EOF
alsa_monitor.enabled = true

alsa_monitor.properties = {
  -- Use ALSA device with highest priority
  ["alsa.reserve"] = true,
  ["alsa.reserve.priority"] = -20,
  ["alsa.reserve.application-name"] = "WirePlumber",
  
  -- ALSA device naming
  ["alsa.midi.node-name"] = "%f",
  ["alsa.midi.node-description"] = "%f",
}

alsa_monitor.rules = {
  {
    matches = {
      {
        -- Match all audio devices
        { "node.name", "matches", "alsa_input.*" },
        { "node.name", "matches", "alsa_output.*" },
      },
    },
    apply_properties = {
      ["node.nick"] = "ALSA Device",
      ["priority.driver"] = 100,
      ["priority.session"] = 100,
    },
  },
}
EOF

sudo tee /etc/wireplumber/main.lua.d/90-enable-all-audio.lua <<EOF
table.insert(alsa_monitor.rules, {
  matches = {
    {
      { "device.name", "matches", "alsa_card.*" },
    },
  },
  apply_properties = {
    ["api.alsa.use-acp"] = true,
    ["api.acp.auto-profile"] = true,
    ["api.acp.auto-port"] = true,
  },
})
EOF

sudo tee /etc/init.d/pipewire <<EOF
#!/sbin/openrc-run
name="PipeWire"
description="PipeWire Media Server"

command="/usr/bin/pipewire"
command_background="yes"
pidfile="/run/pipewire.pid"
supervisor="supervise-daemon"

# Ensure proper environment
export XDG_RUNTIME_DIR="/var/run/pipewire"
export DBUS_SESSION_BUS_ADDRESS="unix:path=/var/run/dbus/system_bus_socket"

# Create necessary directories
start_pre() {
    # Create directory with proper permissions
    checkpath -d -m 0775 -o root:audio /var/run/pipewire
    
    # Verify dbus is running
    if ! service_started dbus; then
        eerror "D-Bus is not running. PipeWire requires D-Bus."
        return 1
    fi
}

# Set proper permissions after PipeWire starts
start_post() {
    # Wait a moment for sockets to be created
    sleep 2
    
    # Set permissions on all PipeWire sockets
    chmod 660 /var/run/pipewire/pipewire-0
    chmod 660 /var/run/pipewire/pipewire-0-manager
    chown root:audio /var/run/pipewire/pipewire-0
    chown root:audio /var/run/pipewire/pipewire-0-manager
}

depend() {
    need dbus
    after alsasound
}
EOF

sudo tee /etc/init.d/wireplumber <<EOF
#!/sbin/openrc-run
name="WirePlumber"
description="PipeWire Session Manager"

command="/usr/bin/wireplumber"
command_background="yes"
pidfile="/run/wireplumber.pid"
supervisor="supervise-daemon"

# Set environment
start_pre() {
    export XDG_RUNTIME_DIR=/var/run/pipewire
    export DBUS_SESSION_BUS_ADDRESS="unix:path=/var/run/dbus/system_bus_socket"
    
    # Verify pipewire is running
    if ! service_started pipewire; then
        eerror "PipeWire is not running. WirePlumber requires PipeWire."
        return 1
    fi
}

depend() {
    need dbus pipewire
    after pipewire
}
EOF

sudo chmod 755 /etc/init.d/pipewire
sudo chmod 755 /etc/init.d/wireplumber

sudo mkdir -p /etc/pipewire/client.conf.d/
sudo tee /etc/pipewire/client.conf.d/99-system-access.conf <<EOF
client.access = {
    # Make it explicitly use the socket in /var/run/pipewire
    path = "/var/run/pipewire/pipewire-0"
    allowed = [ "audio" ]
}
EOF

echo 'export XDG_RUNTIME_DIR=/var/run/pipewire' >> ~/.profile
echo 'export PIPEWIRE_RUNTIME_DIR=/var/run/pipewire' >> ~/.profile

mkdir -p ~/.local/share/
ln -sf /var/run/pipewire ~/.local/share/pipewire

source ~/.profile

sudo rc-update add pipewire default
sudo rc-update add wireplumber default

sudo rc-service pipewire start
sleep 2  # Give PipeWire time to initialize
sudo rc-service wireplumber start

pgrep -a pipewire
pgrep -a wireplumber

pw-cli help
pw-cli list-objects

pw-record test.wav  # Press Ctrl+C after a few seconds to stop

groups $(whoami)

ls -l /var/run/pipewire/

sudo apk add dbus

rc-update add dbus

rc-service dbus start

cat <<EOF | sudo tee /etc/init.d/dbus-user
#!/sbin/openrc-run

command="/usr/bin/dbus-daemon"
pidfile="/run/user/$(id -u)/dbus.pid"
DBUS_USER_DIR="/run/user/$(id -u)"

depend() {
    need localmount
}

start_pre() {
    # Ensure the user runtime directory exists
    if [ ! -d "$DBUS_USER_DIR" ]; then
        mkdir -p "$DBUS_USER_DIR"
        chown $(id -u):$(id -g) "$DBUS_USER_DIR"
        chmod 700 "$DBUS_USER_DIR"
    fi
}

start() {
    ebegin "Starting dbus-user"
    start-stop-daemon --start --background --make-pidfile --pidfile "$pidfile" \
        --exec $command -- --session --address=unix:path=$DBUS_USER_DIR/bus --nofork
    eend $?
}

EOF

sudo chmod +x /etc/init.d/dbus-user
sudo rc-update add dbus-user default
sudo rc-service dbus-user start

echo 'export DBUS_SESSION_BUS_ADDRESS="unix:path=/run/user/$(id -u)/bus"' >> ~/.profile

2025-03-17

Maybe I can use NFC Tags that are attached to the PCs which then point to the related URL?

2025-03-20

I implemented the ALSA installation in an Ansible Playbook and tested it on a different device. After some debugging it worked and the audio could be listened to on the website.

2025-04-15

Today I am going to brainstorm some ideas for the installation setup and going to make some decisions. Currently I am looking at these images as a possible direction.

I showed Katja these images and she recommended to me that I should look up the artist Adam Basanta.

On his Instagram I found the installations Crawler (a model of learning, Invisible Labour and The Unknown Future Rolls Towards Us, which I found aesthetically inspiring.

Inspired by these installation I decided to build mine in a similar fashion with angled iron profiles. Ideally they would also be recycled but for a faster prototyping I decided to buy some new ones.

2025-04-16

It is time to get the phone running! Based on this tutorial I am going to try to install Alpine on it inside of Termux.

As the tutorial recommends to use PRoot Distro, I will try this first.

pkg upgrade -y

pkg install proot-distro

pd install alpine

pd login alpine --fix-low-ports

apk add openrc

apk add openssh

touch /run/openrc/softlevel

rc-service sshd status

rc-update add sshd

rc-service sshd start

pd remove alpine

pkg remove proot-distro

pkg update & pkg upgrade

pkg install wget curl openssh neofetch nmap -y

sshd

passwd

ifconfig

whoami

ssh-copy-id -p 8022 u0_a47@10.10.51.202

ssh -p 8022 u0_a47@10.10.51.202

pkg install ncurses-utils

infocmp -x ghostty | ssh -p 8022 u0_a47@10.10.51.202 -- tic -x -

nano $PREFIX/etc/ssh/sshd_config

pkg update && pkg upgrade

cd $CODE/sys-net-visible/server && GOOS=android GOARCH=arm64 CGO_ENABLED=0 go build -o fairphone-server cmd/api/main.go

# Server folder
mkdir opt
mkdir opt/sys-net-visible-server
mkdir opt/sys-net-visible-server/html

chmod 755 opt/sys-net-visible-server
chmod 755 opt/sys-net-visible-server/html

# Server binary
scp -P 8022 -r $CODE/sys-net-visible/server/fairphone-server u0_a47@10.10.51.202:~/opt/sys-net-visible-server/

# Server config
scp -P 8022 -r $CODE/sys-net-visible/server/fairphone-config.json u0_a47@10.10.51.202:~/opt/sys-net-visible-server/config.json

# Server HTML
scp -P 8022 -r $CODE/sys-net-visible/public/node01/* u0_a47@10.10.51.202:~/opt/sys-net-visible-server/html

cd ~/opt/sys-net-visible-server && ./fairphone-server

pkg install alsa-utils

pkg update 
pkg install pulseaudio

pulseaudio --start --exit-idle-time=-1

pactl load-module module-sles-source

mkdir -p ~/.termux
mkdir -p ~/.termux/boot/

nano $PREFIX/etc/profile.d/boot-pulseaudio.sh

#!/data/data/com.termux/files/usr/bin/sh

# Kill any existing pulseaudio processes
pkill -f pulseaudio

# Clean up any leftover files that might prevent pulseaudio from starting
rm -f ~/.config/pulse/pid
rm -rf /data/data/com.termux/files/usr/tmp/pulse-*

# Start PulseAudio with specific socket path in Termux environment
pulseaudio --start --exit-idle-time=-1 --verbose

# Give PulseAudio a moment to initialize
sleep 5

# Load the SLES source module
pactl load-module module-sles-source

# Verify pulseaudio is running properly
if pactl info >/dev/null 2>&1; then
    echo "PulseAudio setup completed successfully!"
else
    echo "PulseAudio failed to start properly. Please check the logs."
fi

chmod +x $PREFIX/etc/profile.d/boot-pulseaudio.sh

nano ~/.termux/boot/enable-pulseaudio

#!/data/data/com.termux/files/usr/bin/sh
termux-wake-lock
source /data/data/com.termux/files/usr/etc/profile.d/boot-pulseaudio.sh

nano ~/.termux/boot/start-sshd

#!/data/data/com.termux/files/usr/bin/sh
termux-wake-lock
sshd

#!/data/data/com.termux/files/usr/bin/sh
# Keep Termux awake to ensure the server stays running
termux-wake-lock

# Create logs directory if it doesn't exist
mkdir -p ~/logs

# Change to the server directory
cd ~/opt/sys-net-visible-server/

# Start the server in the background with output logging
./server > ~/logs/sys-net-visible-server.log 2>&1 &

# Optionally, write the PID to a file for easier management
echo $! > ~/logs/sys-net-visible-server.pid

# Print a message indicating server has started
echo "Server started with PID $(cat ~/logs/sy-net-visible-server.pid)"

2025-04-17

Today I will try to add an SSH Tunnel to the Phone server.

nano $HOME/.bashrc

VPS_USER="USER" # User on the VPS (e.g., root, admin, tunneluser)
VPS_HOSTNAME="IP_ADDRESS" # IP address or hostname of your VPS
REMOTE_PORT="PORT" # The port on the VPS that will forward to the local port
LOCAL_PORT="8080" # The local port on the Termux device to forward (e.g., the Termux SSH port)
TUNNEL_KEY_PATH="$HOME/.ssh/tunnel_key" # Path for the specific key
SSH_CONFIG_PATH="$HOME/.ssh/config"
# Optional: Define a specific alias/host entry for the VPS connection in SSH config
VPS_HOST_ALIAS="vps-tunnel-host"

source $HOME/.bashrc

pkg update
pkg upgrade -y # Recommended before installing new packages
pkg install -y openssh autossh termux-services termux-api

ssh-keygen -t ed25519 -f "$TUNNEL_KEY_PATH" -N ""

# Ensure private key has correct permissions (usually set by ssh-keygen, but doesn't hurt to check)
chmod 600 "$TUNNEL_KEY_PATH"
chmod 644 "${TUNNEL_KEY_PATH}.pub"

# Create or append to the SSH config file
cat << EOF >> "$SSH_CONFIG_PATH"

Host ${VPS_HOST_ALIAS}
    HostName ${VPS_HOSTNAME}
    User ${VPS_USER}
    Port 22 # Or your VPS SSH port if not standard 22
    IdentityFile ${TUNNEL_KEY_PATH}
    # Optional: Add options for robustness if needed
    # ServerAliveInterval 60
    # ServerAliveCountMax 3
EOF

# Set correct permissions for the config file
chmod 600 "$SSH_CONFIG_PATH"

ssh-copy-id -i "${TUNNEL_KEY_PATH}.pub" -p 722 "${VPS_USER}@${VPS_HOSTNAME}"

mkdir -p "$PREFIX/var/service/reverse-tunnel/log"

# Create the run script
cat << EOF > "$PREFIX/var/service/reverse-tunnel/run"
#!/data/data/com.termux/files/usr/bin/sh
export AUTOSSH_GATETIME=0
export AUTOSSH_POLL=60
# Optional: uncomment the next line to use termux-wake-lock
termux-wake-lock
# Ensure the log directory exists (good practice)
mkdir -p "$PREFIX/var/service/reverse-tunnel/log/main"
# Run autossh, redirecting stderr to stdout for logging
exec autossh -M 0 -N -T -R ${REMOTE_PORT}:localhost:${LOCAL_PORT} ${VPS_HOST_ALIAS} 2>&1
EOF

# Make the run script executable
chmod +x "$PREFIX/var/service/reverse-tunnel/run"

# Create the log run script
cat << EOF > "$PREFIX/var/service/reverse-tunnel/log/run"
#!/data/data/com.termux/files/usr/bin/sh
# Ensure the log directory exists
mkdir -p "$PREFIX/var/service/reverse-tunnel/log/main"
# Run the logger
exec svlogd -tt "$PREFIX/var/service/reverse-tunnel/log/main"
EOF

# Make the log run script executable
chmod +x "$PREFIX/var/service/reverse-tunnel/log/run"

# Enable the service (creates symlinks)
sv-enable reverse-tunnel

# Start the service (or use 'sv up')
sv up reverse-tunnel

# Check the status
sv status reverse-tunnel

# Check the logs (wait a few seconds after starting)
tail -f "$PREFIX/var/service/reverse-tunnel/log/main/current"
# Press Ctrl+C to stop viewing logs

mkdir -p $HOME/.termux/boot

cat << EOF > "$HOME/.termux/boot/start-daemons"
#!/data/data/com.termux/files/usr/bin/sh
termux-wake-lock

# Start the main service manager daemon in the background
# This daemon will automatically start all enabled services (like your autossh one)
termux-services-daemon &

# You generally don't need to explicitly start individual services here if they are enabled.
# The daemon handles it.
EOF

# Make it executable
chmod +x "$HOME/.termux/boot/start-daemons"

mkdir -p $PREFIX/var/service/sshd/log

nano $PREFIX/var/service/sshd/run

#!/data/data/com.termux/files/usr/bin/sh
# -D: Don't detach and become a daemon
# -e: Log to stderr
# sshd will use the configuration from $PREFIX/etc/ssh/sshd_config (default port 8022)
exec sshd -D -e

chmod +x $PREFIX/var/service/sshd/run

nano $PREFIX/var/service/sshd/log/run

#!/data/data/com.termux/files/usr/bin/sh
# Create log directory if it doesn't exist
mkdir -p "$PREFIX/var/service/sshd/log/main"
# Execute svlogd
exec svlogd -tt "$PREFIX/var/service/sshd/log/main"

chmod +x $PREFIX/var/service/sshd/log/run

sv enable sshd

sv status sshd

mkdir -p $PREFIX/var/service/pulseaudio

nano $PREFIX/var/service/pulseaudio/run

#!/data/data/com.termux/files/usr/bin/sh
LOGFILE="$HOME/pulseaudio_service_native_load.log"
echo "--- Starting PulseAudio Service (Explicit Native Module Load): $(date) ---" > $LOGFILE
# Keep foreground, keep disable-shm, reduce verbosity slightly
exec pulseaudio --disable-shm=yes --daemonize=no -v >> $LOGFILE 2>&1

chmod +x $PREFIX/var/service/pulseaudio/run

mkdir -p $PREFIX/var/service/pulseaudio/log/

nano $PREFIX/var/service/pulseaudio/log/run

#!/data/data/com.termux/files/usr/bin/sh
# Set the logging directory for this service
LOG_DIR="$PREFIX/var/log/pulseaudio"
mkdir -p "$LOG_DIR"
# Use svlogd to manage logs, rotating them in the specified directory
exec svlogd "$LOG_DIR"

chmod +x $PREFIX/var/service/pulseaudio/log/run

nano ~/.config/pulse/default.pa

load-module module-sles-source
load-module module-native-protocol-unix

sv enable pulseaudio

sv start pulseaudio

mkdir -p $PREFIX/var/service/web-server

nano $PREFIX/var/service/web-server/run

#!/data/data/com.termux/files/usr/bin/sh

# Change to the server directory
cd ~/opt/sys-net-visible-server/

# Start server
exec ./server

chmod +x $PREFIX/var/service/web-server/run

mkdir -p $PREFIX/var/service/web-server/log/

nano $PREFIX/var/service/web-server/log/run

#!/data/data/com.termux/files/usr/bin/sh
# Set the logging directory for this service
LOG_DIR="$PREFIX/var/service/web-server/log/main"
mkdir -p "$LOG_DIR"
# Run logger
exec svlogd -tt "$LOG_DIR"

chmod +x $PREFIX/var/service/web-server/log/run

sv enable web-server

sv start web-server

sv status $PREFIX/var/service/*

ClientAliveInterval 60  
ClientAliveCountMax 2

2025-04-18

Currently running into the problem that the reverse tunnel to the VPS is not working. I do not know why.

I added the goserver service as a dependency to the reverse-tunnel service in the OpenRC file. Probably the tunnel needs pot 80 to run before it can initialise itself.

depend() {
    need net goserver
    after firewall
}

This does not seem to fix the issue… :/

Apparently the negotiation between the DCHP server and the PCs does take longer in some cases then goserver and reverse-tunnel need to start up. As a quick fix I added a sleep to the pre startup section of the reverse-tunnel service.

start_pre() {
    # Ensure directory exists for pidfile
    checkpath --directory --owner root:root --mode 0755 /run

    sleep 60
}

2025-04-22

Bought the domain visiblesys.net for the work. That will help with writing the URL into your phone while visiting the exhibitions. The subdomains will be accordingly:

• n-1.visiblesys.net
• n-2.visiblesys.net
• n-3.visiblesys.net
• n-4.visiblesys.net

It also thought about creating more specific subdomains that somehow include the title of the website shown on them. I will got with the n-* way for now and the marketing material and maybe implement a redirect if I decide to switch subdomains.

Then I started working on my website for the installation. It will be a documentation of the making of /sys/net/visible, similar to the Request for Comments from the early days of the internet.

For implementing it I want to stay as “web native” as possible, both in terms of style and implementation. Therefore I do not want to use a lot of extra styling or third-party JS libraries. Current sketch looks like this.

After having the first draft done I implemented a pandoc command with custom Lua scripts to convert the logs files from Markdown to a HTML page.

2025-04-25

An idea for a text on the poster:

The installation repurposes three old PCs and an old smartphone as web servers. The physical hardware of the internet - the servers - is staged in the installation. Accordingly, four websites - the digital software of the internet - are presented on the web. The four websites are a text, a game, a sound and a map, each dealing in its own way with the implications and effects of the infrastructure of the internet.

2025-04-27

Finished with layouting, printing and setting up the posters for the exhibition. Also the promotional e-mail is sent and social media posts are posted.

Now back to the phone server. Clemens modified the phones battery with a buck converter so that we can power it continuously. That way the USB-C port can be used for the ethernet connection.

Now I need to redo the setup of the terminal with the new rooted phone.

I updated the android version from the settings of the phone. This disabled the root access of Magisk… -.- So I had to flash the phone again.

Then I followed my earlier logs to setup the phone again.

2025-05-03

Testing the setup for the exhibition at the Oldenburger Computer-Museum.

2025-05-06

Vernissage & Kolloquium for /sys/net/visible @ Oldenburger Computermuseum.

2025-09-24

Got the chance to exhibit /sys/net/visible again during the Transform 2025 – Conference on AI, Art, Design and Society. For this transport I packed everything up into Euroboxes to be a bit faster while unpacking and installing. Installation went pretty seamless and everything was working as expected.

The installation was positioned above the speaker stage on a scaffolding. That was kinda cool. :D