The public IP was previously detected independently in three places,
each contacting a different third party: the Hub (HTTPS echo via
api.ipify.org / ifconfig.me / icanhazip.com on every API call and
background tick), DDNS (myip.opendns.com via OpenDNS), and LiveKit
(embedded STUN). Consolidate into a single detector with one shared
cache so every consumer reads the same value with minimal exposure.
- add modules/core/public-ip.nix: installs /var/lib/sovran/public-ip.py
(pure Python stdlib, no new deps) writing /var/lib/secrets/external-ip
- detection chain (first success wins): explicit pin, fresh cache
(default TTL 300s), STUN binding request over UDP (one packet, no
metadata), DNS myip.opendns.com query, then OPT-IN HTTPS echo
(publicIP.httpsEcho, empty by default — never contacted unless listed)
- privacy: while the cache is fresh zero third parties are contacted;
at most one party learns the IP per refresh interval, via the least
exposing mechanism available
- hub (server.py): _get_external_ip() now reads the shared detector /
cache instead of calling ipify/ifconfig/icanhazip directly
- ddns (njalla.nix): use the shared detector instead of a separate
OpenDNS dig; allow the hardened service to write /var/lib/secrets
- element-calling: livekit-turn-setup falls back to the shared
detector on cold boot; add LiveKit webhooks to lk-jwt-service
(sfu_webhook) so abrupt disconnects are cleaned up immediately;
set LIVEKIT_SANITY_CHECK_INTERVAL_SECONDS=60 as a missed-webhook
guard; drop the dead services.livekit.settings block and set
openFirewall=false (Caddy fronts the SFU; no public 7880/tcp)
- new options: sovran_systemsOS.publicIP.{stunServer,stunPort,
dnsResolver,httpsEcho,cacheTTL}
NixOS already knows whether a reboot is pending: /nix/var/nix/profiles/
system vs /run/current-system. Marker files only the Hub's own updater
wrote desynced for terminal-updated machines (and markers from older
updaters could never clear), pinning the badge on forever. Reconcile
REBOOT_REQUIRED against live state on every read; the stale marker
self-heals to IDLE. The .generation marker write is now informational.
The full-system updater runs as a detached systemd service and can finish
successfully even when the browser loses its status connection. In that
case the update log and status file correctly report REBOOT_REQUIRED, but
the Hub modal can remain on "Updating..." with its controls disabled.
There were four independent ways for the frontend to get stuck:
* update status fetches had no deadline, so a request that stayed pending
never rejected and never advanced the existing failure counter;
* setInterval started async polls without waiting for the previous poll,
allowing slow requests to overlap and responses to arrive out of order;
* each log chunk used textContent +=, replacing the complete and growing
Nix build log every two seconds, which could stall browser rendering and
was especially visible over RDP; and
* page reload, tab resume, and RDP reconnect did not reattach the modal to
the update status persisted by the backend.
This produced a dangerous UX mismatch: the machine had a fully staged
NixOS generation and was ready to reboot, while the Hub continued telling
the user that the update was still running.
Bound status requests with AbortController, prevent overlapping polls, and
replace the endless spinner after sustained failures with an explicit
"Update status unavailable" state and Retry Status action. Reconcile state
immediately on focus, visibility, online, page startup, and before starting
a new update. Use no-store requests and render verbose logs incrementally
with a bounded visible tail while retaining the complete report in memory.
Apply the same timeout and single-flight protection to rebuild polling.
Record the exact generation produced by `nixos-rebuild boot`. The Hub now
keeps REBOOT_REQUIRED visible until that generation matches
/run/current-system, then clears the marker after reboot. For an update
started by an older updater that did not write the marker, recover the
staged generation from the final nixos-rebuild log line. The dashboard
sidebar also distinguishes update-in-progress and restart-required states.
Regression coverage verifies generation marker/log recovery, pre- versus
post-reboot detection, request timeout wiring, single-flight polling,
connection-loss UX, RDP/tab resume reconciliation, bounded log rendering,
page-reload recovery, and JavaScript syntax.
Validation:
* python3 -m unittest discover -s tests -p 'test_*.py' -v (170 passed)
* node --check app/sovran_systemsos_web/static/js/*.js
* python3 -m py_compile for changed Python modules
* git diff --check
A Nix evaluation was not available in the development sandbox; the NixOS
module should still be evaluated and built in CI or on a test machine before
release.
The LND-only rewrite of lndconnect.nix shipped a wrapper Zeus cannot
use: unknown flags (--cert/--macaroon), a non-existent onion path
(free/lnd.onion), and a REST hidden service that collided with LND's
P2P onion. Restore the nix-bitcoin contract — dedicated lnd-rest
onion on port 8080, --nocert over Tor, admin macaroon in the URI —
and only persist a valid lndconnect:// URI for the Hub QR.
- Replace plain-text write of new_root_password in api_security_reset()
with scrypt-hashed storage via _hash_password(), matching how the free
password is already handled.
- Return new_root_password in the API response so the user sees it once
before it is irreversibly hashed on disk.
- Teach _resolve_credential() to detect scrypt hashes and display a
human-readable placeholder instead of raw hex in the Hub credentials UI.
- Harden root-password-setup systemd service: if the secrets file already
contains a hash, skip chpasswd so a manual restart never sets the hash
as the literal login password.
The Caddyfile is generated at runtime by caddy-generate-config.service
from /var/lib/domains/*, but the generator only re-runs when caddy.service
starts fresh. Saving a domain while Caddy is already running therefore
never seats the new virtual host — no proxying and no ACME cert — and the
Hub's reachability check shows a misleading 'ports 80/443' router error
until the next reboot or rebuild.
api_domains_set and api_domains_set_email now restart the generator and
reload Caddy (ExecReload: caddy reload --force, no dropped connections)
right after saving. Entirely skipped when Caddy is inactive — e.g. Node
role before its first domain-based service is enabled — because the
rebuild that enables the service starts caddy.service for the first time,
runs the generator first (requiredBy), and seats the already-saved domain
on its own. Best-effort throughout: a domain save never fails because of
a Caddy reload issue.
Co-authored-by: arena-agent <297053741+arena-agent@users.noreply.github.com>
- The Hub header badge could render 'vdev' because the runtime
/etc/nixos/VERSION lookup fell back to the literal string 'dev' when
the file was missing (e.g. dev/test environments, or before the
Nix-generated config carried a version at all).
- modules/core/sovran-hub.nix now reads the repo's VERSION file at
Nix eval time and bakes a real semantic version (sovran_version)
into the generated config.json and a VERSION file shipped with the
package, so the Hub always has a solid value to display.
- server.py's _get_sovran_version() now reads that baked-in
sovran_version first, and explicitly rejects a literal 'dev' value
from any of its file-based fallbacks so the badge never shows
'vdev' again.
- templates/index.html + header.css: wrapped the title and the
version badge in a '.title-group' column so the version badge sits
directly underneath 'Sovran_SystemsOS Hub', left-aligned with it.
Co-authored-by: arena-agent <297053741+arena-agent@users.noreply.github.com>
- Add release-stable.sh script with automatic tagging, CHANGELOG updates,
GitHub/Gitea release creation, and VERSION file management
- Update iso/common.nix to include version from VERSION file in ISO filename
- Add VERSION file (current: 1.0.3)
- Polish OS version badge in Sovran Hub header (top-right)
- Update README.md download link to versioned ISO
- Update CHANGELOG.md with existing tags (v1.0.0 – v1.0.3)
Co-authored-by: arena-agent <297053741+arena-agent@users.noreply.github.com>
Previously only bitcoind.service showed a version on the tile.
Now every service with a parsable Nix store path gets its version
displayed (electrs, btcpayserver, lnd, rtl, mempool, vaultwarden,
matrix-synapse, caddy, tor, etc.).
The existing _get_service_version() helper reads systemctl show
<unit> --property=ExecStart and extracts the version from the Nix
store path. Results are cached for 5 minutes per unit.
Co-authored-by: arena-agent <297053741+arena-agent@users.noreply.github.com>
Naming: user-facing 'Wallet Connections' -> 'Lightning Wallet Connections'
across the Hub, feature registry, tile, and NixOS modules. Internal ids
(nwc-wallets, albyhub.service, /api/nwc/*) are unchanged.
UX: the service-detail modal put status, domain diagnostics, router ports,
the enable/disable toggle, restart, the liquidity guide and the whole wallet
manager in one cramped scrolling column. For this feature the modal is now
980px wide and split into two tabs:
- Wallets: wallet grid, create/share/verify flows, collapsible liquidity guide
- Service & Setup: description, status, domain checklist, ports, enable, restart
A status dot and domain chip sit in the tab bar so state is visible from both
tabs, and the modal opens on Setup when the service is off or the Lightning
Address domain is unconfigured. Wallet cards gain a balance chip, pending
badge, a prominent address row, and separated destructive actions.
Non-NWC services keep the original single-column layout and width.
Co-authored-by: arena-agent <297053741+arena-agent@users.noreply.github.com>
Each NWC wallet connection gets its own shareable LNURL QR so the owner
can receive payments from anyone without creating invoices (the core
LNURL value proposition: a static QR that never expires).
Hub UI (service detail modal):
- New 'Share QR' button on every wallet connection card
- Dedicated share view: large QR, Lightning Address + copy, raw bech32
LNURL + copy
- Download PNG (websites/social), download SVG (vector, print-sharp at
any size), and a Print button that opens a print-ready payment card
which auto-invokes the browser print dialog
Backend (FastAPI, session-authenticated):
- GET /api/nwc/wallets/{id}/lnurl -> alias, lightning_address, LNURL
- GET /api/nwc/wallets/{id}/lnurl-qr.png (?download=1, ?scale=)
- GET /api/nwc/wallets/{id}/lnurl-qr.svg (?download=1)
- GET /api/nwc/wallets/{id}/lnurl-qr/print (print-ready HTML card)
- Stdlib bech32 encoder (BIP-173, verified against the official LUD-01
golden vector); LNURL is uppercased for QR alphanumeric-mode density
per LUD-01 guidance; QR rendered with the existing qrencode package
- Resolves wallets by app id or pubkey; 404 unknown wallet, 503 when no
Lightning domain is configured; all routes return 401 unauthenticated
No new system dependencies.
Co-authored-by: arena-agent <297053741+arena-agent@users.noreply.github.com>
The Element Call tile still appended a synthetic 'Step 4: Router Setup
Needed' to the domain diagnostic checklist, and every port table carried a
'Sovran_SystemsOS Status' column with Ready / Not ready yet verdicts.
Both were misleading: port forwarding happens on the router, which this
computer cannot inspect. A local ss/firewall probe can neither prove nor
disprove that forwarding works — and the LiveKit TURN relay range binds on
demand, so it reported 'Not ready yet' even on a perfectly working system.
- server.py: add ROUTER_FORWARD_ONLY_UNITS ({livekit.service}); skip the
local probe for those units, drop the step-4 append, replace extra_ports
with router_ports (no status field), and exclude router-only ports from
both tile health and /api/ports/health so they can't raise false alarms.
- helpers.js: new shared renderPortForwardGuideHtml() — one intro naming the
internal IP, explicit instructions (same internal/external port, match the
protocol, use port-range fields for 30000-40000), a colour-coded
TCP / UDP / TCP+UDP badge per row, and a closing note that the only real
test is loading the service from a phone on mobile data.
- features.js: enable-time modal uses the shared guide and now always lists
every port to forward (the old local pre-filter hid ports the user still
had to open).
- service-detail.js: tile port section uses the same guide; the SSH/non-domain
branch keeps a small 'not open on this computer yet' hint, which is a real
local fact, separate from router forwarding.
- onboarding.js: step 3 router note reworded to match (same number for
internal/external, notes that Element Call adds UDP ports).
- domain-setup.css: styles for the protocol badges and instruction list.
Co-authored-by: arena-agent <297053741+arena-agent@users.noreply.github.com>
- Onboarding: remove the redundant/error-prone 'Router Setup' step (5 steps -> 4).
A compact 80/443 (+22 SSH) note now lives inside Domain Configuration,
and the Element Call ports are only shown at the moment they matter:
when enabling the feature, and afterwards on the service tile.
- Onboarding step 3: fix domain prefill bug (API returns {domains: {...}}),
make /api/network fetch best-effort so it can never block the step.
- Enable-time port modal: streamline copy (one intro + table + pointer to
the tile's live status view).
- Element Call tile detail: replace 5 repetitive prose blocks with 2 compact
notes around the live-status port table.
- Njal.la DDNS: run njalla.sh immediately when a DDNS-backed feature is
enabled (previously only ran on domain save or the 15-min cron tick).
- Harden njalla.sh handling: create the base script (shebang + IP lookup)
if missing before appending curl lines; invoke via bash explicitly.
Co-authored-by: arena-agent <297053741+arena-agent@users.noreply.github.com>
- nwc_hub_manager.py: Fix Authorization header to use real ****** (was hardcoded to literal asterisks due to display redaction)
- server.py: Use exc.args[0] instead of str(exc) in NWC error handlers to prevent CodeQL stack-trace taint flow to HTTP responses; update albyhub.service description key
- Add nwc_hub_manager.py: AlbyHubManager with real Alby Hub API (setup, auth, CRUD, drain, delete, invoice)
- Add nwc_lnurl_service.py: dedicated loopback LNURL service on port 8181
- server.py: remove JSON scaffolding (state.json, fake invoice generator, fake NWC URI, LNURL routes); replace with real manager calls; update service maps to albyhub.service; remove LNURL auth-exempt paths
- nwc_wallet_cli.py: rewrite to use real AlbyHubManager instead of JSON state
- modules/nwc-wallets.nix: replace with albyhub user/service, nwc-lnurl service, LND macaroon, unlock-password generation
- modules/core/caddy.nix: proxy LNURL routes to port 8181 (dedicated service) instead of 8937 (Hub)
- modules/core/sovran-hub.nix: service tile points to albyhub.service
- docs/wallet-connections.md: document real architecture, Alby Hub pin/patches, backup sensitivity
- test_wallet_connections.py: replace scaffolding tests with 54 real manager tests using mocked Alby Hub
naturallaw777
naturallaw777
Sovran Systems
Arena Agent
copilot-swe-agent[bot]andnaturallaw777
Contributor
