dearsky/numa
1
0
Fork 0
Code Issues 10 Pull Requests 1 Actions Packages Projects Releases 25 Wiki Activity
288 Commits 4 Branches 25 Tags
6fc59a7bd5aa2b4651313a2517d1525fc670a727
Commit Graph

6 Commits

Author SHA1 Message Date
Razvan Dimescu
7efac85836 feat: wire-level forwarding, cache, request hedging, and DoH keepalive 
Wire-level forwarding path skips DnsPacket parse/serialize on the hot
path. Cache stores raw wire bytes with pre-scanned TTL offsets — patches
ID + TTLs in-place on lookup instead of cloning parsed packets.

Request hedging (Dean & Barroso "Tail at Scale") fires a second
parallel request after a configurable delay (default 10ms) when
the primary upstream stalls. DoH keepalive loop prevents idle
HTTP/2 + TLS connection teardown.

Recursive resolver now hedges across multiple NS addresses and
caches NS delegation records to skip TLD re-queries.

Integration test harness polls /blocking/stats instead of fixed
sleep, eliminating the blocklist-download race condition.
 2026-04-12 18:39:48 +03:00
Razvan Dimescu
7d6b0ed568 feat: DoH server endpoint + DoT enabled by default (#79) 
* chore: document multi-forwarder and cache warming in config and README

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* feat: DNS-over-HTTPS server endpoint (RFC 8484)

Serve DoH at POST /dns-query on the existing HTTPS proxy (port 443).
Automatically enabled when proxy TLS is active — no config needed.
Also fix zone map priority so local zones override RFC 6762 .local
special-use handling.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* style: cargo fmt

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* chore: remove GoatCounter analytics from site

GoatCounter domains (goatcounter.com, gc.zgo.at) are blocked by
Hagezi Pro, which is Numa's default blocklist. A DNS privacy tool
should not embed analytics that its own resolver blocks.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* feat: enable DoT listener by default

DoT now starts automatically with `sudo numa`, matching the proxy and
DoH which are already on by default. The self-signed CA infrastructure
is shared with the proxy, so there is no additional setup. This makes
`numa setup-phone` work out of the box.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
 2026-04-11 04:06:17 +03:00
Razvan Dimescu
6887c8e02e refactor: move data_dir override from env var to [server] TOML field 
Reverts the NUMA_DATA_DIR env var added in the previous commit and
replaces it with a [server] data_dir TOML field. Numa already has a
well-developed config system; adding a parallel env-var mechanism
for a single knob was wrong.

The principle: TOML is for application behavior configuration. Env
vars are for bootstrap values (HOME, SUDO_USER to discover paths
before config loads) and standard ecosystem conventions (RUST_LOG).
data_dir is neither — it's an app knob, so it belongs in the TOML.

Changes:
- lib.rs::data_dir() reverts to the platform-specific fallback only
- config.rs adds `data_dir: Option<PathBuf>` to ServerConfig
- main.rs resolves config.server.data_dir with fallback to
  numa::data_dir() and passes it to build_tls_config, then stores the
  resolved path on ctx.data_dir for downstream consumers
- tls.rs::build_tls_config takes `data_dir: &Path` as an explicit
  parameter instead of calling crate::data_dir() behind the caller's
  back. regenerate_tls and dot.rs self_signed_tls now pass
  &ctx.data_dir, honoring whatever path the config resolved to
- tests/integration.sh Suite 6 uses `data_dir = "$NUMA_DATA"` in its
  test TOML instead of the NUMA_DATA_DIR env var prefix
- numa.toml gains a commented-out data_dir example

No behavior change for existing production deployments (the default
path is unchanged). Test harness is now fully config-driven, and
containerized deploys can override data_dir via mount+config without
needing env var injection.

127/127 unit tests pass, Suite 6 passes end-to-end.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
 2026-04-08 02:53:43 +03:00
Razvan Dimescu
7f52bd8a32 test: Suite 6 — proxy + DoT coexistence, NUMA_DATA_DIR override 
Adds integration test coverage for the realistic production shape
where both the HTTPS proxy and DoT are enabled simultaneously. This
was previously untested — every existing suite had either one or the
other, so the interaction path was implicit.

What Suite 6 verifies:
- Both listeners bind without panic
- DoT still resolves queries with the proxy enabled
- Proxy HTTPS handshake still works with DoT enabled
- Both certs validate against the same shared CA

To run non-root, adds a NUMA_DATA_DIR env var override to data_dir()
that lets callers point the CA/cert storage at any writable path.
Useful beyond tests: containerized deployments, CI runners, dev
testing without sudo. The fallback is the existing platform-specific
path (unix: /usr/local/var/numa, windows: %PROGRAMDATA%\numa).

Suite 6 sets NUMA_DATA_DIR=/tmp/numa-integration-data before
starting numa, then trusts the generated CA at $NUMA_DATA_DIR/ca.pem
for both kdig (DoT query) and openssl s_client (HTTPS proxy
handshake) verification.

All 6 suites, 32 checks, run non-root and pass locally.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
 2026-04-08 02:53:43 +03:00
Razvan Dimescu
c98e6c3ea9 fix: install rustls crypto provider when loading user DoT cert 
Adds tests/integration.sh Suite 5 (DoT via kdig + openssl) and
fixes a startup panic caught by it.

Bug: when [dot] cert_path/key_path was set AND [proxy] was disabled,
numa panicked on the first DoT handshake with "Could not
automatically determine the process-level CryptoProvider from Rustls
crate features". In normal deployments the proxy's build_tls_config
installs the default provider as a side effect, masking the missing
call in dot.rs::load_tls_config. Disable the proxy and the panic
surfaces. Fix: call
rustls::crypto::ring::default_provider().install_default() at the
top of load_tls_config (no-op if already installed).

Suite 5 exercises:
- DoT listener binds on configured port
- Resolves a local zone A record over TLS (kdig +tls)
- Persistent connection reuse (kdig +keepopen, 3 queries, 1 handshake)
- ALPN "dot" negotiation (openssl s_client -alpn dot)
- ALPN mismatch rejected with no_application_protocol (openssl -alpn h2)

Uses a pre-generated cert at /tmp so the test runs non-root.
Skips gracefully if kdig or openssl aren't installed.

Also: Dockerfile now EXPOSE 853/tcp so docker run -p 853:853 works
out of the box when users enable DoT.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
 2026-04-08 02:53:43 +03:00
Razvan Dimescu
a84f2e7f1d feat: recursive DNS + DNSSEC + TCP fallback (#17) 
* feat: recursive resolution + full DNSSEC validation

Numa becomes a true DNS resolver — resolves from root nameservers
with complete DNSSEC chain-of-trust verification.

Recursive resolution:
- Iterative RFC 1034 from configurable root hints (13 default)
- CNAME chasing (depth 8), referral following (depth 10)
- A+AAAA glue extraction, IPv6 nameserver support
- TLD priming: NS + DS + DNSKEY for 34 gTLDs + EU ccTLDs
- Config: mode = "recursive" in [upstream], root_hints, prime_tlds

DNSSEC (all 4 phases):
- EDNS0 OPT pseudo-record (DO bit, 1232 payload per DNS Flag Day 2020)
- DNSKEY, DS, RRSIG, NSEC, NSEC3 record types with wire read/write
- Signature verification via ring: RSA/SHA-256, ECDSA P-256, Ed25519
- Chain-of-trust: zone DNSKEY → parent DS → root KSK (key tag 20326)
- DNSKEY RRset self-signature verification (RRSIG(DNSKEY) by KSK)
- RRSIG expiration/inception time validation
- NSEC: NXDOMAIN gap proofs, NODATA type absence, wildcard denial
- NSEC3: SHA-1 iterated hashing, closest encloser proof, hash range
- Authority RRSIG verification for denial proofs
- Config: [dnssec] enabled/strict (default false, opt-in)
- AD bit on Secure, SERVFAIL on Bogus+strict
- DnssecStatus cached per entry, ValidationStats logging

Performance:
- TLD chain pre-warmed on startup (root DNSKEY + TLD DS/DNSKEY)
- Referral DS piggybacking from authority sections
- DNSKEY prefetch before validation loop
- Cold-cache validation: ~1 DNSKEY fetch (down from 5)
- Benchmarks: RSA 10.9µs, ECDSA 174ns, DS verify 257ns

Also:
- write_qname fix for root domain "." (was producing malformed queries)
- write_record_header() dedup, write_bytes() bulk writes
- DnsRecord::domain() + query_type() accessors
- UpstreamMode enum, DEFAULT_EDNS_PAYLOAD const
- Real glue TTL (was hardcoded 3600)
- DNSSEC restricted to recursive mode only

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* feat: TCP fallback, query minimization, UDP auto-disable

Transport resilience for restrictive networks (ISPs blocking UDP:53):
- DNS-over-TCP fallback: UDP fail/truncation → automatic TCP retry
- UDP auto-disable: after 3 consecutive failures, switch to TCP-first
- IPv6 → TCP directly (UDP socket binds 0.0.0.0, can't reach IPv6)
- Network change resets UDP detection for re-probing
- Root hint rotation in TLD priming

Privacy:
- RFC 7816 query minimization: root servers see TLD only, not full name

Code quality:
- Merged find_starting_ns + find_starting_zone → find_closest_ns
- Extracted resolve_ns_addrs_from_glue shared helper
- Removed overall timeout wrapper (per-hop timeouts sufficient)
- forward_tcp for DNS-over-TCP (RFC 1035 §4.2.2)

Testing:
- Mock TCP-only DNS server for fallback tests (no network needed)
- tcp_fallback_resolves_when_udp_blocked
- tcp_only_iterative_resolution
- tcp_fallback_handles_nxdomain
- udp_auto_disable_resets
- Integration test suite (4 suites, 51 tests)
- Network probe script (tests/network-probe.sh)

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* feat: DNSSEC verified badge in dashboard query log

- Add dnssec field to QueryLogEntry, track validation status per query
- DnssecStatus::as_str() for API serialization
- Dashboard shows green checkmark next to DNSSEC-verified responses
- Blog post: add "How keys get there" section, transport resilience section,
  trim code blocks, update What's Next

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* fix: use SVG shield for DNSSEC badge, update blog HTML

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* fix: NS cache lookup from authorities, UDP re-probe, shield alignment

- find_closest_ns checks authorities (not just answers) for NS records,
  fixing TLD priming cache misses that caused redundant root queries
- Periodic UDP re-probe every 5min when disabled — re-enables UDP
  after switching from a restrictive network to an open one
- Dashboard DNSSEC shield uses fixed-width container for alignment
- Blog post: tuck key-tag into trust anchor paragraph

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* fix: TCP single-write, mock server consistency, integration tests

- TCP single-write fix: combine length prefix + message to avoid split
  segments that Microsoft/Azure DNS servers reject
- Mock server (spawn_tcp_dns_server) updated to use single-write too
- Tests: forward_tcp_wire_format, forward_tcp_single_segment_write
- Integration: real-server checks for Microsoft/Office/Azure domains

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* feat: recursive bar in dashboard, special-use domain interception

Dashboard:
- Add Recursive bar to resolution paths chart (cyan, distinct from Override)
- Add RECURSIVE path tag style in query log

Special-use domains (RFC 6761/6303/8880/9462):
- .localhost → 127.0.0.1 (RFC 6761)
- Private reverse PTR (10.x, 192.168.x, 172.16-31.x) → NXDOMAIN
- _dns.resolver.arpa (DDR) → NXDOMAIN
- ipv4only.arpa (NAT64) → 192.0.0.170/171
- mDNS service discovery for private ranges → NXDOMAIN

Eliminates ~900ms SERVFAILs for macOS system queries that were
hitting root servers unnecessarily.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* chore: move generated blog HTML to site/blog/posts/, gitignore

- Generated HTML now in site/blog/posts/ (gitignored)
- CI workflow runs pandoc + make blog before deploy
- Updated all internal blog links to /blog/posts/ path
- blog/*.md remains the source of truth

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* fix: review feedback — memory ordering, RRSIG time, NS resolution

- Ordering::Relaxed → Acquire/Release for UDP_DISABLED/UDP_FAILURES
  (ARM correctness for cross-thread coordination)
- RRSIG time validation: serial number arithmetic (RFC 4034 §3.1.5)
  + 300s clock skew fudge factor (matches BIND)
- resolve_ns_addrs_from_glue collects addresses from ALL NS names,
  not just the first with glue (improves failover)
- is_special_use_domain: eliminate 16 format! allocations per
  .in-addr.arpa query (parse octet instead)

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* feat: API endpoint tests, coverage target

- 8 new axum handler tests: health, stats, query-log, overrides CRUD,
  cache, blocking stats, services CRUD, dashboard HTML
- Tests use tower::oneshot — no network, no server startup
- test_ctx() builds minimal ServerCtx for isolated testing
- `make coverage` target (cargo-tarpaulin), separate from `make all`
- 82 total tests (was 74)

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
 2026-03-28 04:03:47 +02:00