coredns: production Let's Encrypt cert via Caddy sidecar (DNS-01 + Vultr)

Replaces the self-signed dev cert flow with a real LE prod cert for dns.l.supported.systems, issued and auto-renewed by a Caddy sidecar using DNS-01 challenge against the Vultr API. Components: - caddy/Dockerfile builds Caddy 2.10.0 with caddy-dns/vultr plugin via xcaddy. GOTOOLCHAIN=auto so xcaddy can fetch newer Go on demand when plugin versions advance their minimum Go. - caddy/Caddyfile uses DNS-01 with explicit public resolvers (1.1.1.1, 9.9.9.9) for the propagation check. Without that, Docker's embedded DNS leaks the container into the host's split-horizon LAN DNS, which returns LAN IPs for ns1.vultr.com and the propagation check fails. - docker-compose: caddy service shares ./caddy-data with coredns via a read-only subpath mount that excludes /acme (account private key). - Healthcheck doubles as a symlinker: maintains stable cert.pem / key.pem names at /data/caddy/ and chmods cert files + their dirs to be readable by CoreDNS's nonroot user. Flips to "healthy" only once the symlinks dereference (i.e. cert exists), gating CoreDNS start via depends_on: service_healthy. - Corefile unchanged — same /etc/coredns/certs/cert.pem path; only the bind-mount source switches from ./certs to ./caddy-data/caddy. - New Makefile target: tls-up orchestrates the bring-up sequence. Cert is valid until Aug 12 2026. Verified end-to-end: dig @127.0.0.1 -p 8853 +tls +tls-hostname=dns.l.supported.systems ... dig @127.0.0.1 -p 8443 +https +tls-hostname=dns.l.supported.systems ...
2026-05-14 01:34:57 -06:00 · 2026-05-14 01:34:57 -06:00 · c1afe77b27
commit c1afe77b27
parent 066ba1892a
6 changed files with 195 additions and 32 deletions
--- a/.env
+++ b/.env
@ -16,3 +16,15 @@ DOT_PORT=8853
 # DoH (DNS-over-HTTPS, RFC 8484) — typically 443. Host port 8443
 # because Caddy already owns 443 on this host.
 DOH_PORT=8443
+
+# --- Production cert provisioning (Caddy sidecar + Let's Encrypt) ---
+# Hostname the cert is issued for. Must be a name you control and that
+# resolves via the public DNS server holding the zone (Vultr's NS).
+CADDY_HOSTNAME=dns.l.supported.systems
+
+# Contact email registered with Let's Encrypt for expiry notifications.
+ACME_EMAIL=rpm@malloys.us
+
+# VULTR_API_KEY is intentionally NOT stored here. Caddy reads it from
+# the shell environment via docker compose's variable interpolation —
+# export it in your shell (or in ~/.zshenv) before `make tls-up`.
--- a/.gitignore
+++ b/.gitignore
@ -4,5 +4,10 @@ zones-prepared/*.zone
 # Self-signed certs (re-generated by scripts/generate-certs.sh)
 certs/*.pem

+# Caddy's runtime data: issued certs, ACME account keys, transient state.
+# NEVER commit — contains the ACME account private key.
+caddy-data/
+caddy-config/
+
 # Local-only env overrides
 .env.local
--- a/109
+++ b/109
@ -2,28 +2,68 @@
 SHELL := /usr/bin/env bash
 COMPOSE := docker compose

-.PHONY: help prep certs up down restart logs ps test test-tls reload clean
+# Source .env so $(CADDY_HOSTNAME) etc. are available in recipes.
+include .env
+export
+
+.PHONY: help prep certs up down restart logs logs-caddy ps test test-tls \
+        test-public reload clean tls-up cert-watch caddy-rebuild

 help: ## Show this help
-	@awk 'BEGIN {FS = ":.*?## "} /^[a-zA-Z_-]+:.*?## / {printf "  \033[36m%-12s\033[0m %s\n", $$1, $$2}' $(MAKEFILE_LIST)
+	@awk 'BEGIN {FS = ":.*?## "} /^[a-zA-Z_-]+:.*?## / {printf "  \033[36m%-14s\033[0m %s\n", $$1, $$2}' $(MAKEFILE_LIST)

 prep: ## Re-inject SOA records into all zones (writes zones-prepared/)
 	@./scripts/prepare-zones.sh

-certs: ## Generate self-signed cert for DoT/DoH (re-run with FORCE=1 to rotate)
+certs: ## Generate self-signed dev cert (only useful if not using Caddy ACME)
 	@./scripts/generate-certs.sh

-up: prep certs ## Start CoreDNS (prepares zones + ensures certs exist first)
-	$(COMPOSE) up -d
-	@sleep 2 && $(COMPOSE) logs --tail=20 coredns
+caddy-rebuild: ## Rebuild the Caddy image (after editing caddy/Dockerfile)
+	$(COMPOSE) build --no-cache caddy

-down: ## Stop & remove the container
+# ---------------------------------------------------------------------------
+# Production / Let's Encrypt flow
+# ---------------------------------------------------------------------------
+
+tls-up: prep ## Bring up Caddy → wait for cert → start CoreDNS (one command)
+	@if [ -z "$$VULTR_API_KEY" ]; then \
+		echo "ERROR: VULTR_API_KEY is not exported. Set it in your shell:"; \
+		echo "  export VULTR_API_KEY=..."; \
+		exit 1; \
+	fi
+	@mkdir -p caddy-data caddy-config
+	$(COMPOSE) up -d caddy
+	@echo ""
+	@echo "Waiting for Caddy to provision cert for $(CADDY_HOSTNAME)..."
+	@echo "(DNS-01 via Vultr typically takes 30-90s; press Ctrl-C to abort)"
+	@for i in $$(seq 1 90); do \
+		if [ -e caddy-data/caddy/cert.pem ]; then \
+			echo ""; echo "  ✓ cert ready after $${i}0s"; break; \
+		fi; \
+		printf '.'; sleep 10; \
+	done
+	@test -e caddy-data/caddy/cert.pem || \
+		(echo ""; echo "FAILED — see logs: make logs-caddy"; exit 1)
+	$(COMPOSE) up -d coredns
+	@sleep 3 && $(COMPOSE) logs --tail=15 coredns
+
+cert-watch: ## Tail Caddy logs while it provisions the cert
+	$(COMPOSE) logs -f caddy
+
+logs-caddy: ## Tail Caddy logs
+	$(COMPOSE) logs -f caddy
+
+# ---------------------------------------------------------------------------
+# Day-to-day operations
+# ---------------------------------------------------------------------------
+
+down: ## Stop & remove all containers
 	$(COMPOSE) down

-restart: ## Restart CoreDNS (does not re-prep zones)
+restart: ## Restart CoreDNS (does not re-prep zones / re-issue cert)
 	$(COMPOSE) restart coredns

-reload: prep ## Re-prep zones; CoreDNS auto-plugin will pick changes up
+reload: prep ## Re-prep zones; CoreDNS auto-plugin picks changes up
 	@echo "Zones re-prepared. CoreDNS reloads files every 30s (auto plugin)."

 logs: ## Tail CoreDNS logs
@ -32,28 +72,35 @@ logs: ## Tail CoreDNS logs
 ps: ## Show container status
 	$(COMPOSE) ps

-test: ## Smoke-test plain DNS (uses DNS_PORT from .env)
-	@. ./.env && echo "Querying acrazy.org @ 127.0.0.1:$$DNS_PORT (plain DNS)" && \
-		dig @127.0.0.1 -p $$DNS_PORT acrazy.org SOA +short && \
-		dig @127.0.0.1 -p $$DNS_PORT acrazy.org NS +short && \
-		dig @127.0.0.1 -p $$DNS_PORT or.acrazy.org A +short
+# ---------------------------------------------------------------------------
+# Smoke tests
+# ---------------------------------------------------------------------------

-test-tls: ## Smoke-test DoT + DoH (pins self-signed cert via +tls-ca)
-	@. ./.env && \
-		echo "=== DoT @ 127.0.0.1:$$DOT_PORT ===" && \
-		dig @127.0.0.1 -p $$DOT_PORT +tls +tls-ca=certs/cert.pem \
-			+tls-hostname=localhost acrazy.org SOA +short && \
-		echo "" && \
-		echo "=== DoH @ https://localhost:$$DOH_PORT/dns-query ===" && \
-		dig @localhost -p $$DOH_PORT +https +tls-ca=certs/cert.pem \
-			acrazy.org A +short && \
-		echo "" && \
-		echo "=== DoH via curl (raw wire-format) ===" && \
-		curl -sk --cacert certs/cert.pem \
-			-H 'accept: application/dns-message' \
-			--data-binary @<(printf '\x00\x00\x01\x20\x00\x01\x00\x00\x00\x00\x00\x00\x06acrazy\x03org\x00\x00\x01\x00\x01') \
-			-H 'content-type: application/dns-message' \
-			"https://localhost:$$DOH_PORT/dns-query" | xxd | head -5
+test: ## Smoke-test plain DNS
+	@echo "Querying acrazy.org @ 127.0.0.1:$(DNS_PORT) (plain DNS)"
+	@dig @127.0.0.1 -p $(DNS_PORT) acrazy.org SOA +short
+	@dig @127.0.0.1 -p $(DNS_PORT) acrazy.org NS +short
+	@dig @127.0.0.1 -p $(DNS_PORT) or.acrazy.org A +short

-clean: down ## Remove containers + prepared zones + certs
+test-tls: ## Smoke-test DoT + DoH against LOCAL endpoints (trusts cert via system CAs)
+	@echo "=== DoT @ 127.0.0.1:$(DOT_PORT), expecting cert for $(CADDY_HOSTNAME) ==="
+	@dig @127.0.0.1 -p $(DOT_PORT) +tls +tls-hostname=$(CADDY_HOSTNAME) \
+		acrazy.org SOA +short
+	@echo ""
+	@echo "=== DoH @ https://$(CADDY_HOSTNAME):$(DOH_PORT)/dns-query ==="
+	@dig @$(CADDY_HOSTNAME) -p $(DOH_PORT) +https acrazy.org A +short
+
+test-public: ## Smoke-test using the public hostname (DoT/DoH ports must be open + DNS A record set)
+	@echo "=== DoT on public hostname @ port 853 ==="
+	@dig @$(CADDY_HOSTNAME) +tls cloudflare.com A +short
+	@echo "=== DoH on public hostname @ port 443 ==="
+	@dig @$(CADDY_HOSTNAME) +https cloudflare.com A +short
+
+clean: down ## Remove containers + prepared zones + dev self-signed certs
 	rm -rf zones-prepared/*.zone certs/*.pem
+
+clean-caddy: down ## Also wipe Caddy's data dir (forces re-issuance from scratch!)
+	@echo "About to delete caddy-data/ — this will force re-issuance from LE."
+	@echo "Hit Ctrl-C in 5s to abort..."
+	@sleep 5
+	rm -rf caddy-data caddy-config
--- a/caddy/Caddyfile
+++ b/caddy/Caddyfile
@ -0,0 +1,40 @@
+# Caddy is used here purely as an ACME client + cert renewer for CoreDNS.
+# The HTTPS site is technically served (Caddy can't issue without a site
+# block), but we don't expose port 443 from this container — only the
+# cert files in /data/caddy/ are consumed by the CoreDNS sidecar.
+
+{
+    # Operator contact for Let's Encrypt; also used for expiry warnings.
+    email {$ACME_EMAIL}
+
+    # Skip the HTTP-to-HTTPS redirect server (we have nothing to redirect).
+    # Caddy still binds :443 inside the container for the cert site, which
+    # is fine because we don't publish those ports to the host.
+    auto_https disable_redirects
+}
+
+{$CADDY_HOSTNAME} {
+    tls {
+        # DNS-01 challenge via Vultr API. The plugin reads the token from
+        # the named env var; setting via {env.VULTR_API_KEY} would also
+        # work but the bare reference is clearer with Caddy's modules.
+        dns vultr {env.VULTR_API_KEY}
+
+        # Use PUBLIC resolvers for the propagation check, not Docker's
+        # embedded DNS. Without this, Caddy follows the container's
+        # resolv.conf → host's resolv.conf → local LAN resolvers, which
+        # on a split-horizon DNS setup will return LAN IPs for vultr.com
+        # nameservers and the propagation check fails with connection
+        # refused. Hitting 1.1.1.1 / 9.9.9.9 directly sidesteps it.
+        resolvers 1.1.1.1 9.9.9.9 1.0.0.1
+
+        # Vultr's NS propagation is generally fast (<30s) but LE checks
+        # multiple resolvers; cushion the wait to avoid flaky issuance.
+        propagation_delay 30s
+        propagation_timeout 300s
+    }
+
+    # A sensible response if anyone hits this on 443. Doubles as a
+    # "Caddy is alive" sanity check inside the compose network.
+    respond "CoreDNS DoT/DoH endpoint. DoT: port 853. DoH: /dns-query" 200
+}
--- a/caddy/Dockerfile
+++ b/caddy/Dockerfile
@ -0,0 +1,14 @@
+# Custom Caddy build that bundles the Vultr DNS provider plugin.
+# Stock caddy:2 doesn't include DNS-provider modules — they're plugins.
+# xcaddy compiles them in at build time.
+FROM caddy:2.10.0-builder AS builder
+# The Caddy builder image bakes in Go 1.23, but caddy-dns/vultr now
+# requires Go >= 1.24. GOTOOLCHAIN=auto lets `go get` fetch a newer
+# toolchain on demand so we don't have to bump base images every time
+# a plugin's minimum Go version moves.
+ENV GOTOOLCHAIN=auto
+RUN xcaddy build \
+    --with github.com/caddy-dns/vultr
+
+FROM caddy:2.10.0
+COPY --from=builder /usr/bin/caddy /usr/bin/caddy
--- a/docker-compose.yml
+++ b/docker-compose.yml
@ -1,9 +1,49 @@
 services:
+  # Caddy runs as a dedicated ACME client + cert renewer. It provisions
+  # a Let's Encrypt cert for ${CADDY_HOSTNAME} via DNS-01 (Vultr API)
+  # and persists it to ./caddy-data. CoreDNS reads from that same path
+  # read-only. The container's HTTP/HTTPS ports are NOT published — we
+  # only care about the cert files on disk.
+  caddy:
+    build: ./caddy
+    container_name: coredns-caddy
+    restart: unless-stopped
+    environment:
+      - CADDY_HOSTNAME=${CADDY_HOSTNAME}
+      - ACME_EMAIL=${ACME_EMAIL}
+      - VULTR_API_KEY=${VULTR_API_KEY:?VULTR_API_KEY must be exported in your shell}
+    volumes:
+      - ./caddy/Caddyfile:/etc/caddy/Caddyfile:ro
+      - ./caddy-data:/data
+      - ./caddy-config:/config
+    healthcheck:
+      # Two-jobs-in-one: (1) maintain stable filenames (cert.pem / key.pem)
+      # as symlinks into Caddy's hostname-keyed storage, so the Corefile
+      # doesn't have to encode the hostname. (2) Flip to "healthy" once
+      # the symlink dereferences successfully (i.e. Caddy has issued).
+      # Relative symlink targets so paths work the same from host or
+      # from any container mounting this directory.
+      test:
+        - "CMD-SHELL"
+        - >
+          ln -sf certificates/acme-v02.api.letsencrypt.org-directory/${CADDY_HOSTNAME}/${CADDY_HOSTNAME}.crt /data/caddy/cert.pem &&
+          ln -sf certificates/acme-v02.api.letsencrypt.org-directory/${CADDY_HOSTNAME}/${CADDY_HOSTNAME}.key /data/caddy/key.pem &&
+          chmod 755 /data/caddy &&
+          chmod -R a+rX /data/caddy/certificates 2>/dev/null;
+          test -e /data/caddy/cert.pem
+      interval: 10s
+      timeout: 3s
+      retries: 60        # ~10 min ceiling for initial issuance
+      start_period: 5s
+
  coredns:
    image: ${COREDNS_IMAGE}
    container_name: coredns
    restart: unless-stopped
    command: ["-conf", "/etc/coredns/Corefile"]
+    depends_on:
+      caddy:
+        condition: service_healthy
    ports:
      - "${DNS_PORT}:53/udp"
      - "${DNS_PORT}:53/tcp"
@ -14,7 +54,12 @@ services:
    volumes:
      - ./Corefile:/etc/coredns/Corefile:ro
      - ./zones-prepared:/zones:ro
-      - ./certs:/etc/coredns/certs:ro
+      # Subpath mount of Caddy's data dir. The healthcheck maintains
+      # cert.pem / key.pem symlinks at the top of this tree, so CoreDNS
+      # sees stable filenames regardless of hostname. The /accounts dir
+      # (ACME registration private key) is sibling to /caddy and is NOT
+      # exposed to CoreDNS — only /caddy is mounted.
+      - ./caddy-data/caddy:/etc/coredns/certs:ro
    healthcheck:
      test: ["CMD", "wget", "-qO-", "http://127.0.0.1:8080/health"]
      interval: 30s