Root cause: credentials.json had flat-format test data from old code,
incompatible with current W3C VerifiableCredential struct. Parse error
was hidden by error sanitization.
Fix: cleared old test data. VC flow now works bidirectionally:
- .198: 3/3 issue + 3/3 verify
- .228: issue + verify work (rate-limited during repeated testing)
- Both nodes: list-credentials returns correct counts
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Both nodes rebooted simultaneously. .228 SSH in 115s, .198 in ~5min.
Both healthy. Federation re-established — 2 peers synced.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Session tokens get invalidated when backend restarts. Moving auth
inside the iteration loop ensures each iteration gets a fresh session.
Also fix grep -c arithmetic syntax error for nostr-provider check.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Root cause: sd_notify::notify(true, ...) cleared NOTIFY_SOCKET env var,
so watchdog pings never reached systemd. Backend killed every 60s.
Fixes:
- Change sd_notify::notify first param to false (keep socket)
- Increase WatchdogSec from 60 to 300 (5min) for crash recovery
- Add TimeoutStartSec=300 for slow container startups
- Adjust watchdog ping interval to 120s
This was causing 47 restarts/day on .198 and blocking REBOOT-03,
FLEET-03, FLEET-04, VC-04.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Removed 54 unused/dangling images from .228.
50% total image disk reduction (freed 26.96GB).
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Fix: bash parameter splitting caused {} to break into body JSON.
Changed rpc() to declare params separately.
Removed set -e to allow individual test failures.
FLEET-02: .228 passes 30/30 (3 iterations) — all features validated.
FLEET-03: .198 blocked — backend instability, 15/28 pass.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Simplify DHT encoding: use JSON instead of DNS packets (drop simple-dns)
- Fix mainline crate API: SigningKey takes 32 bytes, get_mutable returns Result
- Add missing dht_did field to IdentityRecord constructor
- Store DID Document as JSON in DHT (DNS encoding deferred)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- DHT Identity card with blue status indicator
- "Publish to DHT" button calls identity.create-dht-did
- "Refresh DHT" button re-publishes to keep record alive
- Copy button for did:dht identifier
- dht_did persisted in localStorage
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- federation.list-nodes now includes vc_verified: bool per node
- True when a non-revoked FederationTrustCredential exists for the peer DID
- Integrates with VC-02's automatic VC issuance on federation join
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Issue W3C VC (type FederationTrustCredential) when joining federation
- Claims: federationPeer=true, establishedAt=timestamp
- Signed with node Ed25519 identity key
- Runs in background task (non-blocking)
- Stored via credentials system for later verification
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Add dht_did field to IdentityRecord (optional, serde-compatible)
- Add prefer_dht_did param to identity.issue-credential RPC
- When true and dht_did is set, uses did:dht as VC issuer
- Credential system already format-agnostic for any DID type
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- PERF-01: Move crash recovery to background tokio task so health
endpoint is available immediately on startup
- PERF-04: Add ResponseCache with 5s TTL for system.stats and
federation.list-nodes. Reduces CPU for frequent polling.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Crash recovery (check_for_crash + recover_containers +
start_stopped_containers) now runs in a background tokio task.
The health endpoint is available immediately on startup instead of
blocking for 260+ seconds while containers restart sequentially.
This directly fixes the .198 boot recovery timeout issue where the
backend took 260s to become healthy after restart.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Initial load: 110KB gzipped (index.js). All views code-split.
Total: 312KB gzipped across all chunks. No optimization needed.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- TAP format, takes target IP + --iterations N
- Checks: health, memory, disk, containers, federation, DWN,
identity, NIP-07, backup create/verify/delete
- Exit 0 = production ready
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Add tier: "" to all AppMetadata match arms (was missing from 30+ arms)
- Use std::thread::available_parallelism() instead of num_cpus crate
- Remove unused num_cpus dependency
- Fix unused variable warning in health_monitor.rs
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Show 'core' (orange) and 'recommended' (blue) badges next to app titles
- getAppTier() classifies apps matching backend get_app_tier()
- Global .tier-badge, .tier-badge-core, .tier-badge-recommended CSS classes
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
.198 crash recovery takes >120s for 34 containers. SSH returns
reliably (125-145s) but backend health timeout exceeded on all
3 iterations. Needs CONT-02 deployment and/or increased timeout.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
get_app_tier() classifies all apps:
- core: Bitcoin, LND, Electrs, Mempool, BTCPay, DWN, FileBrowser
- recommended: Fedimint, Grafana, Vaultwarden, Kuma, SearXNG, etc.
- optional: everything else
Tier field added to Manifest struct (data_model.rs) and exposed
via WebSocket package data for frontend tier badges.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Per-container RAM/CPU/disk measurements from .228 baseline.
Three app tiers: Core (2.6GB), Recommended (+880MB), Optional (+2-5GB).
Four hardware tiers with cost estimates.
10K user distribution projection.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
DEPLOY-02: --canary flag deploys to both then verifies .198 health
DEPLOY-03: Pre-deploy rollback backup (binary + web-ui) to
/opt/archipelago/rollback/. Auto-rollback on post-deploy health failure.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Add swap creation to first-boot-containers.sh
- Size: 50% of RAM (min 2GB, max 8GB)
- Creates /swapfile, adds to /etc/fstab for persistence
- Runs before container creation to prevent OOM during startup
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Shows target, mode, files to sync, build steps, and deploy scope
without executing any changes. Works with --live, --both, etc.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
MemoryTracker in health_monitor.rs tracks per-container RSS every 5 min.
Warns when a container's memory grows >50% over tracking period.
Parses podman stats output (GiB/MiB/KiB formats).
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
MEM-01: OOM kill detection via dmesg checks every 5 minutes
MEM-03: Disk growth rate tracking (288 samples over 24h), warns at >1GB/day
MEM-04: Systemd watchdog (WatchdogSec=60, sd_notify::Watchdog every 30s)
Service Type=notify for proper startup notification
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Creates scripts/test-reboot-survival.sh with TAP format output.
Records pre-reboot containers, reboots node, waits for SSH + health,
verifies container count/state/health. 6 checks per iteration.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Add US-10 backup/restore test section to test-cross-node.sh
- Test cycle: create → list → verify → delete, 10 iterations × 2 nodes
- Increase backup.create rate limit from 3/600 to 10/600 (still conservative)
- Increase backup.restore rate limit from 2/600 to 5/600
- Clean up 21K+ stale DWN test messages on both servers
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Make dwn.sync endpoint async: spawns background task, returns immediately
- Add 90s overall timeout to sync_with_peers via tokio::time::timeout
- Deduplicate peer onion addresses before syncing
- Batch message pushes (50 per request) instead of one-at-a-time over Tor
- Add 15s connect_timeout to Tor SOCKS5 client
- Cap local message query to 200 messages per sync
- Fix DWN HTTP handler to process ALL messages in batch (was only first)
- Add recordId deduplication in handler to prevent duplicate imports
- Update test script to poll dwn.status for sync completion
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Fixed ssh_sudo in US-07 section where chown ran without sudo because
&& in the command broke the sudo pipe. With set -e, this silently killed
the script. Wrapped compound commands in sudo bash -c to keep everything
under sudo. All file sharing tests pass bidirectionally over Tor.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
