wrolpi/repair.sh
lrnselfreliance d701846254
Some checks are pending
CodeQL / Analyze (push) Waiting to run
Controller Tests / controller-tests (push) Waiting to run
Jest Tests / jest-tests (push) Waiting to run
Hotspot captive portal: welcome page with the WROLPi's address (#709)
* Hotspot captive portal: welcome page with the WROLPi's address

Devices that join the hotspot probe a fixed plain-HTTP URL to check for
Internet access.  NetworkManager's shared-mode dnsmasq now resolves those
probe hostnames (and `wrolpi`) to the hotspot address via an include the
Controller writes when the hotspot starts, the probes land on the Controller
(port 80), and are redirected to /portal: "This is your WROLPi, its address is
10.42.0.1, open WROLPi / open the Controller".

Only the probe hostnames are hijacked so the hotspot keeps forwarding
everything else when the WROLPi has an uplink.  The include is corrected and
the connection re-activated once if NetworkManager picked a non-default
subnet.  Writing the include is best-effort; the hotspot starts regardless.

`hotspot.captive_portal` (controller.yaml, default true) toggles the probe
redirects and the include; like the other hotspot settings it applies on the
next hotspot start, so saving while the hotspot is up reports
restart_required and both Controller UIs say so.  The status endpoint gains
the hotspot address and portal URL, shown in both UIs.  repair.sh only
creates the include directory so a disabled setting survives upgrades.
Native only: Docker has no hotspot.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>

* captive portal: address review

* React: re-read the hotspot address whenever the HotspotToggle starts or
  stops the hotspot, matching the fallback UI's refresh-on-status.
* A WPA3 start that reverts to WPA2 leaves the hotspot up, so it now gets
  the captive portal address correction too.
* The corrective `nmcli connection up` logs a warning on a non-zero exit;
  dnsmasq would otherwise keep the old address silently.
* HotspotStatusResponse.captive_portal defaults to true like everywhere else.
* Trim design narrative from the module docstring and React comments.
* The include-before-activate test now observes the include at the moment
  of the `nmcli ... hotspot` call instead of inferring from call order.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>

* captive portal: never forward any record type for probe hosts

`address=` only synthesizes A/AAAA answers.  iOS also asks for the HTTPS
record type; dnsmasq forwarded that upstream, Apple's answer aliased to
captive.g.aaplimg.com, and the phone fetched the genuine "Success" from
Apple instead of the portal.  Each probe host now also gets `local=`, so
dnsmasq answers every type itself, and the CDN alias is a probe host too.

Observed with dnsmasq query logging on a hotspot with an eth0 uplink.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>

* captive portal: let acknowledged clients finish so the OS offers "Done"

A phone's sign-in sheet only shows "Done" once the OS re-probes and gets
the genuine success reply.  The Controller now remembers each client (by
hotspot IP) that has loaded the portal page and answers its later probes
with the reply each OS expects (Apple "Success" HTML, Android 204, Windows
"Microsoft Connect Test", ...).  The set is cleared when the hotspot
starts so a new session shows the page again.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>

* captive portal: acknowledge on Continue, not on view

iOS fetches the portal page in the background before opening the sheet,
then re-probes after the sheet's own navigation.  Acknowledging on view
made that re-probe return "Success" as the sheet's content.  Phones only
re-check the network (and offer "Done") after a full-page navigation, so
the welcome page now has a Continue button leading to /portal/continue,
which is the only place a client is acknowledged.  The Continue page says
to tap Done and links to WROLPi and the Controller.

Also hijack clients3.google.com (older Android probe host).

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>

* help: bump wrolpi-help to the captive portal and Speed Test docs

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>

* captive portal: comments describe the Continue design; clear acks on stop

Three comments still described acknowledging on view, which c72a7f7b
removed.  Acknowledged clients are now also cleared when the hotspot
stops, so a lease reused in a later session sees the portal page.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>

---------

Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
2026-09-13 20:10:02 -06:00

217 lines
8.7 KiB
Bash
Executable file

#!/usr/bin/env bash
# This script will attempt to repair a WROLPi installation. It will not use internet.
cd /opt/wrolpi || (echo "Cannot repair. /opt/wrolpi does not exist" && exit 1)
# Re-execute this script if it wasn't called with sudo.
if [ $EUID != 0 ]; then
sudo "$0" "$@"
exit $?
fi
rpi=false
if (grep 'Raspberry Pi' /proc/cpuinfo >/dev/null); then
rpi=true
fi
set -e
set -x
# Stop services if they are running.
systemctl stop wrolpi-controller.service || :
systemctl stop wrolpi-api.service || :
systemctl stop wrolpi-app.service || :
systemctl stop wrolpi-kiwix.service || :
systemctl stop caddy || :
# Clear Python bytecode cache from virtual environments to fix potential corruption.
echo "Clearing Python bytecode cache..."
find /opt/wrolpi/venv -type d -name __pycache__ -exec rm -rf {} + 2>/dev/null || :
find /opt/wrolpi/venv -name "*.pyc" -delete 2>/dev/null || :
find /opt/wrolpi/controller/venv -type d -name __pycache__ -exec rm -rf {} + 2>/dev/null || :
find /opt/wrolpi/controller/venv -name "*.pyc" -delete 2>/dev/null || :
find /opt/wrolpi-help/venv -type d -name __pycache__ -exec rm -rf {} + 2>/dev/null || :
find /opt/wrolpi-help/venv -name "*.pyc" -delete 2>/dev/null || :
# Create the WROLPi user
grep wrolpi: /etc/passwd || useradd -md /home/wrolpi wrolpi -s "$(command -v bash)"
# Reset any inadvertent changes to the WROLPi repo. Restore ownership in case this repair script fails.
git config --global --add safe.directory /opt/wrolpi
git reset HEAD --hard
chown -R wrolpi:wrolpi /opt/wrolpi
# Scripts are run by systemd services (e.g. wrolpi-cert-renew.service) which fail with
# "Permission denied" if the executable bit was lost.
chmod +x /opt/wrolpi/*.sh /opt/wrolpi/scripts/*.sh
# Copy configs to system.
mkdir -p /etc/caddy
cp /opt/wrolpi/etc/raspberrypios/Caddyfile /etc/caddy/Caddyfile
mkdir -p /var/www
cp /opt/wrolpi/etc/raspberrypios/50x.html /var/www/50x.html
# The Controller writes the hotspot captive-portal dnsmasq include here when the hotspot starts.
# Never copy a file into it: the include only exists while the setting is enabled.
mkdir -p /etc/NetworkManager/dnsmasq-shared.d
# Generate certificate for HTTPS.
/opt/wrolpi/scripts/generate_certificates.sh
# Install map fonts from blob (needed for MapLibre map labels).
MAP_FONTS_BLOB=/opt/wrolpi-blobs/map-fonts.tar.gz
if [ -f "${MAP_FONTS_BLOB}" ]; then
mkdir -p /opt/wrolpi/modules/map/static/fonts
tar -xzf "${MAP_FONTS_BLOB}" -C /opt/wrolpi/modules/map/static/
fi
# Start Caddy quickly so user can access Controller from React UI
systemctl start caddy || :
# WROLPi needs a few privileged commands.
cp /opt/wrolpi/etc/raspberrypios/90-wrolpi /etc/sudoers.d/90-wrolpi
chmod 0440 /etc/sudoers.d/90-wrolpi
# Verify this new file is valid.
visudo -c -f /etc/sudoers.d/90-wrolpi
# Update the Portable bootstrap on boxes installed from the Live USB; it is
# otherwise frozen at image-build time.
LIVE_INCLUDES=/opt/wrolpi/debian-live-config/config/includes.chroot
if [ -f /usr/local/sbin/wrolpi-bootstrap.sh ]; then
cp "${LIVE_INCLUDES}/usr/local/sbin/wrolpi-bootstrap.sh" /usr/local/sbin/wrolpi-bootstrap.sh
chmod 0755 /usr/local/sbin/wrolpi-bootstrap.sh
if [ -f /etc/systemd/system/wrolpi-bootstrap.service ]; then
cp "${LIVE_INCLUDES}/etc/systemd/system/wrolpi-bootstrap.service" /etc/systemd/system/wrolpi-bootstrap.service
fi
fi
# Install the systemd services
cp /opt/wrolpi/etc/raspberrypios/wrolpi*.service /etc/systemd/system/
cp /opt/wrolpi/etc/raspberrypios/wrolpi-*.timer /etc/systemd/system/
cp /opt/wrolpi/etc/raspberrypios/wrolpi.target /etc/systemd/system/
systemctl enable wrolpi-controller.service
systemctl enable wrolpi-api.service
systemctl enable wrolpi-app.service
systemctl enable wrolpi-kiwix.service
systemctl enable wrolpi-help.service
systemctl enable wrolpi-mounts.service
systemctl enable wrolpi-cert-renew.timer
# Repair Controller (offline-safe - no pip install)
repair_controller() {
echo "Checking WROLPi Controller..."
# Check venv exists
if [ ! -f /opt/wrolpi/controller/venv/bin/python ]; then
echo "ERROR: Controller venv missing!"
echo "Run install.sh or upgrade.sh (requires internet) to reinstall."
return 1
fi
# Check venv is functional
if ! /opt/wrolpi/controller/venv/bin/python --version > /dev/null 2>&1; then
echo "ERROR: Controller venv is corrupted!"
echo "Run install.sh or upgrade.sh (requires internet) to reinstall."
return 1
fi
# Check systemd service is installed
if [ ! -f /etc/systemd/system/wrolpi-controller.service ]; then
echo "Controller systemd service not installed, copying..."
cp /opt/wrolpi/etc/raspberrypios/wrolpi-controller.service /etc/systemd/system/
systemctl daemon-reload
fi
# Ensure service is enabled
if ! systemctl is-enabled wrolpi-controller > /dev/null 2>&1; then
echo "Enabling Controller service..."
systemctl enable wrolpi-controller
fi
echo "Controller repair completed"
}
# Run Controller repair if venv exists (skip on fresh install)
if [ -d /opt/wrolpi/controller/venv ]; then
repair_controller || echo "Controller repair failed, continuing..."
fi
# Start Controller so user can monitor repair in UI.
systemctl start wrolpi-controller
/usr/bin/systemctl daemon-reload
# WROLPi needs a few privileged commands.
cp /opt/wrolpi/etc/raspberrypios/90-wrolpi /etc/sudoers.d/90-wrolpi
chmod 0440 /etc/sudoers.d/90-wrolpi
# Verify this new file is valid.
visudo -c -f /etc/sudoers.d/90-wrolpi
# WROLPi uses SQLite inside the media directory; the API creates/migrates the database at
# startup. Never fatal: the media drive may not be mounted yet.
(cd /opt/wrolpi && /opt/wrolpi/venv/bin/python3 /opt/wrolpi/main.py db upgrade) || \
echo "WARNING: db upgrade failed (media directory unavailable?); the API will retry at startup."
# Create the media directory. This should be mounted by the maintainer.
[ -d /media/wrolpi ] || mkdir /media/wrolpi
# Create config directory if external drive is mounted, and is empty.
if grep -qs /media/wrolpi /proc/mounts && [ -z "$(ls -A /media/wrolpi)" ] && [ ! -d /media/wrolpi/config ]; then
mkdir /media/wrolpi/config
fi
# Build the frontend app -- only if build/ is missing or the app source changed.
# The shipped image already contains a current build/, so on a healthy install
# (including a weak Pi) this is a no-op; it only rebuilds after an upgrade that
# changed app/, or if build/ went missing. Stays offline: it never npm-installs.
/opt/wrolpi/scripts/build_app.sh
# Change owner of the media directory, ignore any errors because the drive may be fat/exfat/etc.
chown wrolpi:wrolpi /media/wrolpi 2>/dev/null || echo "Ignoring failure to change media directory permissions."
# Ensure the config directory is owned by wrolpi so the API can write config files and the SQLite DB.
[ -d /media/wrolpi/config ] && chown -R wrolpi:wrolpi /media/wrolpi/config 2>/dev/null || :
# Remove immutable flag from blob files before chown (may not exist or may not be set).
chattr -i /opt/wrolpi-blobs/* 2>/dev/null || :
# Ship committed blobs (e.g. the default Zim served by Kiwix when the media dir
# has no working zims) into /opt/wrolpi-blobs. repair.sh runs on every install
# and upgrade, so this keeps the blobs current there just like the image builds
# do. It runs after the immutable flag is cleared above (so existing blobs can
# be overwritten) and before immutability is re-applied below.
mkdir -p /opt/wrolpi-blobs
if [ -d /opt/wrolpi/wrolpi/blobs ] && ls /opt/wrolpi/wrolpi/blobs/* >/dev/null 2>&1; then
cp /opt/wrolpi/wrolpi/blobs/* /opt/wrolpi-blobs/
fi
chown -R wrolpi:wrolpi /home/wrolpi /opt/wrolpi*
# Protect blob files from accidental modification or deletion.
# These files are critical for initializing WROLPi.
if [ -d /opt/wrolpi-blobs ] && ls /opt/wrolpi-blobs/* >/dev/null 2>&1; then
chown -R root:root /opt/wrolpi-blobs
chmod 444 /opt/wrolpi-blobs/*
chattr +i /opt/wrolpi-blobs/*
fi
# Copy MOTD once the repair has been successful.
cp /opt/wrolpi/etc/raspberrypios/motd/30-wrolpi.motd /etc/update-motd.d/30-wrolpi
chmod +x /etc/update-motd.d/*
# Copy desktop shortcuts to existing users (new users get them from /etc/skel).
if [[ ${rpi} == true ]]; then
for user_home in /home/wrolpi /home/pi; do
if [ -d "${user_home}/Desktop" ]; then
cp /opt/wrolpi/etc/raspberrypios/*.desktop "${user_home}/Desktop/"
chown -R "$(basename ${user_home})":"$(basename ${user_home})" "${user_home}/Desktop/"
fi
done
# Update skeleton for future users.
mkdir -p /etc/skel/Desktop
cp /opt/wrolpi/etc/raspberrypios/*.desktop /etc/skel/Desktop/
fi
systemctl restart wrolpi-help
systemctl start wrolpi.target
set +x
echo "Repair has completed. You may run /opt/wrolpi/help.sh to check system status."