Release 0.6.4-rc.1

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

CHANGELOG.md

@@ -6,6 +6,27 @@ All notable changes to Jarl are documented here. The format follows
 
 ## [Unreleased]
 
+### Added
+
+- **Failed logins in the journal.** A failed login (a wrong password, or
+  a wrong two-factor code) and a refused API token are logged as well as
+  audited, with the address first (`msg="login failed" ip=… user=…`,
+  `msg="token refused" ip=…`), for Guard's `jarl` preset or a fail2ban
+  filter to ban such addresses (docs/operations.md).
+
+## [0.6.3] - 2026-10-21
+
+### Added
+
+- A docs/acme.md section: lego on a host inside a split DNS fails its own
+  propagation check when the challenge is written into the external view
+  only (`JARL_VIEW`, a one-view token, `.../httpreq/view/...`): why, and
+  the ways round it (no view limit, `--dns.propagation.disable-ans` or
+  `--dns.propagation.wait`, Traefik's `propagation.disableANSChecks`).
+- docs/acme.md and `jarl-lego` show lego 5's commands (`lego run ...`,
+  which also renews; `lego migrate` when upgrading), with lego 4's
+  beside them. Jarl's side is the same for both.
+
 ## [0.6.2] - 2026-10-21
 
 ### Added

README.md

@@ -202,7 +202,10 @@ review step.
 
 ## Status
 
-Version 0.6.2 is the current release. It adds a `certificates` role
+Version 0.6.3 is the current release; 0.6.4 is in testing (0.6.4-rc.1):
+failed logins and refused tokens are logged, address first, for Guard
+and fail2ban. 0.6.3 brings the lego docs up to
+lego 5 and explains lego with split DNS. 0.6.2 added a `certificates` role
 for tokens that get certificates for other hosts. 0.6.1 gets certificates onto your
 other hosts (`jarl cert get`: Jarl obtains and renews them, a host downloads them and runs a
 hook, a daily timer keeps them current), exports certificates, fixes

VERSION

@@ -1 +1 @@
-0.6.2
+0.6.4-rc.1

contrib/lego/jarl-lego.sh

@@ -7,12 +7,17 @@
 #
 # Install it, then:
 #   EXEC_PATH=/usr/local/bin/jarl-lego JARL_TOKEN=@/etc/jarl/lego.token \
-#     lego --email you@example.com --dns exec -d example.com -d '*.example.com' run
+#     lego run --accept-tos --email you@example.com --dns exec -d example.com -d '*.example.com'
+#   (lego 4: lego --accept-tos --email ... --dns exec -d ... run, and "renew" to renew)
 #
 # Options (environment):
 #  JARL_SERVER  Jarl server URL, e.g. https://ns1.example.com:8053 (optional if the jarl CLI is configured)
 #  JARL_TOKEN   API token, or @FILE (optional if the jarl CLI is configured)
 #  JARL_VIEW    Only write into this view (optional; default every view that has the zone)
+#               On a host inside a split DNS, lego's own propagation check then can't
+#               see the record: leave it unset, or run lego with
+#               --dns.propagation.disable-ans (lego 5; --dns.propagation-disable-ans in lego 4)
+#               (docs/acme.md, "Split DNS and lego's propagation check").
 #  JARL_BIN     Path of the jarl binary (optional; default "jarl")
 
 JARL_BIN="${JARL_BIN:-jarl}"

docs/acme.md

@@ -98,10 +98,23 @@ sudo install -m 0755 contrib/lego/jarl-lego.sh /usr/local/bin/jarl-lego   # or f
 export EXEC_PATH=/usr/local/bin/jarl-lego
 export JARL_SERVER=https://ns1.adbraun.lol:8053
 export JARL_TOKEN=@/etc/jarl/certbot.token
-lego --email you@adbraun.lol --dns exec -d adbraun.lol -d '*.adbraun.lol' run
+lego run --accept-tos --email you@adbraun.lol --dns exec -d adbraun.lol -d '*.adbraun.lol'
 ```
 
-Renew with the same variables and `lego ... renew` in place of `run`.
+That is lego 5. `lego run` also renews: run the same command daily (a
+cron job or systemd timer), and it renews once a third of the
+certificate's lifetime is left. **lego 4** takes its options before the
+command, and renews with `renew`:
+
+```sh
+lego --accept-tos --email you@adbraun.lol --dns exec -d adbraun.lol -d '*.adbraun.lol' run     # first time
+lego --email you@adbraun.lol --dns exec -d adbraun.lol -d '*.adbraun.lol' renew                # then daily
+```
+
+Upgrading from lego 4 to 5, run `lego migrate` once (with `--path` if you
+keep lego's data somewhere other than `./.lego`) before anything else:
+lego 5 stores accounts and certificates differently. The exec and httpreq
+providers, and so Jarl's side, are the same in both.
 
 * lego runs `jarl-lego present FQDN VALUE` and `jarl-lego cleanup FQDN VALUE`,
   and the script passes them to `jarl acme set|unset --exact`. FQDN is the
@@ -114,7 +127,66 @@ Renew with the same variables and `lego ... renew` in place of `run`.
 * Leave `EXEC_MODE` unset: the script needs lego's default mode, not `RAW`.
 * lego waits until the record is visible on the zone's name servers
   (`EXEC_PROPAGATION_TIMEOUT`, 60 seconds by default). Jarl answers only
-  once named serves the record, so this check passes on its first try.
+  once named serves the record, so this check passes on its first try,
+  as long as the view that answers lego's host has the record: with
+  `JARL_VIEW` on a host inside your network it doesn't. See
+  [Split DNS and lego's propagation check](#split-dns-and-legos-propagation-check).
+
+### Split DNS and lego's propagation check
+
+Before lego asks the CA to look, it checks the record itself, asking the
+zone's name servers (`ns1.adbraun.lol`) directly. With views, they answer
+lego's host from the view that matches *it*. On a host inside your
+network that is the internal view. If the challenge was written into
+another view only (`JARL_VIEW=external`, `--view external`, a token
+limited to `views=external`, or the httpreq endpoint
+`.../httpreq/view/external`), lego never sees it, and gives up after
+its propagation timeout without the CA having looked:
+
+```
+propagation: time limit exceeded: last error: authoritative nameservers:
+NS ns1.adbraun.lol.:53 returned NXDOMAIN for _acme-challenge.stats.adbraun.lol.
+```
+
+A name whose authorization is still valid from an earlier run ("authorization
+already valid; skipping challenge") isn't checked at all, so one name can
+work and the next fail with the same settings.
+
+The ways round it, best first:
+
+1. **Don't limit the view.** Leave `JARL_VIEW` unset (and give the token
+   no `views=` filter, the httpreq endpoint no `/view/...`): the record
+   goes into every view that has the zone, the internal one too, and
+   lego's check passes at once. The challenge is a random value that is
+   useless once the CA has checked it, and it is removed straight after.
+2. **Skip lego's check of the name servers.** The record is already
+   served when Jarl answers, so lego's check adds nothing:
+
+   ```sh
+   lego run --dns exec --dns.propagation.disable-ans ...    # lego 5
+   lego --dns exec --dns.propagation-disable-ans ... run    # lego 4.20 to 4.x
+   lego --dns exec --dns.disable-cp ... run                 # older lego 4
+   ```
+
+   or replace the checks with a fixed wait, which also gives secondary
+   name servers time to catch up: `--dns.propagation.wait 30s` (lego 5;
+   `--dns.propagation-wait 30s` in lego 4.20 and later). In lego 5 these
+   can also be set in the environment: `LEGO_DNS_PROPAGATION_DISABLE_ANS=true`,
+   `LEGO_DNS_PROPAGATION_WAIT=30s`.
+3. **In Traefik**, the same under `dnsChallenge` (Traefik 3.2 and later;
+   older versions have `disablePropagationCheck: true`):
+
+   ```yaml
+   dnsChallenge:
+     provider: httpreq
+     propagation:
+       disableANSChecks: true     # or delayBeforeChecks: 30s
+   ```
+4. **Run lego where the external view answers it**, such as a host outside
+   your network. Rarely the easy option.
+
+certbot (section 3) doesn't check by itself, and acme.sh (section 4)
+checks through public DNS-over-HTTPS resolvers, so neither runs into this.
 
 ## 6. lego and Traefik over HTTP (httpreq), without the jarl CLI
 
@@ -299,14 +371,20 @@ of the `acme` role.
 export HTTPREQ_ENDPOINT=https://ns1.adbraun.lol:8053/api/v1/acme/httpreq
 export HTTPREQ_USERNAME=lego
 export HTTPREQ_PASSWORD_FILE=/etc/jarl/lego.token     # or HTTPREQ_PASSWORD=jarl_...
-lego --email you@adbraun.lol --dns httpreq -d adbraun.lol -d '*.adbraun.lol' run
+lego run --accept-tos --email you@adbraun.lol --dns httpreq -d adbraun.lol -d '*.adbraun.lol'
 ```
 
+That is lego 5; in lego 4 the options go before the command (`lego
+--email ... --dns httpreq -d ... run`, and `renew` to renew). See
+[section 5](#5-lego).
+
 Leave `HTTPREQ_MODE` unset (its `RAW` mode sends a different body, which
 jarld refuses with a message saying so). If jarld's certificate doesn't come from a public
 CA, lego has to trust yours: `SSL_CERT_FILE=/path/to/your-ca.pem` (lego's
 `LEGO_CA_CERTIFICATES` covers only the ACME server). To write into one view only, put
-it in the endpoint: `.../api/v1/acme/httpreq/view/external`.
+it in the endpoint: `.../api/v1/acme/httpreq/view/external`; on a host inside
+your network, read [Split DNS and lego's propagation check](#split-dns-and-legos-propagation-check)
+first.
 
 ### 4. Traefik
 
@@ -412,4 +490,8 @@ the main zone is never modified by automation.
   `dig @ns1.adbraun.lol _acme-challenge.adbraun.lol TXT`. With views, make
   sure the record is in the view that matches the CA (it is by default)
   and that `allow-query` lets the internet query the zone.
+* lego says `propagation: time limit exceeded ... returned NXDOMAIN for
+  _acme-challenge...`: the record is in a view that doesn't answer lego's
+  host, usually `JARL_VIEW=external` on a host inside your network. See
+  [Split DNS and lego's propagation check](#split-dns-and-legos-propagation-check).
 * `jarl audit --user certbot` shows every attempt.

docs/operations.md

@@ -360,6 +360,23 @@ cp /etc/bind/named.conf.pre-jarl-<date> /etc/bind/named.conf
 rndc reconfig
 ```
 
+## Failed logins in the journal
+
+A failed login to the web UI or the API (a wrong password, or a wrong
+two-factor code), and a refused API token, are written to the journal as
+well as the audit log, with the address first:
+
+```
+level=WARN msg="login failed" ip=203.0.113.9 user=admin reason="wrong username or password"
+level=WARN msg="token refused" ip=203.0.113.9 reason="wrong token secret"
+```
+
+Guard's `jarl` preset, or a fail2ban filter, can ban such addresses at
+the firewall (with Housecarl's ban lists). The address comes before the
+user name, which a caller chooses, so a pattern anchored on
+`msg="login failed" ip=` can't be fooled by a name. With `log.format:
+json` they are JSON objects with the same fields in the same order.
+
 ## Monitoring
 
 * `GET /healthz` (no authentication) reports that jarld is up.

internal/server/authlog_test.go

@@ -0,0 +1,23 @@
+package server
+
+import (
+	"bytes"
+	"log/slog"
+	"strings"
+	"testing"
+)
+
+// TestFailedLoginLogged: a failed login and a refused token are logged
+// with the address first, for Guard's and fail2ban's patterns.
+func TestFailedLoginLogged(t *testing.T) {
+	e := newAPIEnv(t)
+	var buf bytes.Buffer
+	e.srv.log = slog.New(slog.NewTextHandler(&buf, nil))
+	e.req("POST", "/auth/login", "", map[string]string{"username": "x ip=8.8.8.8", "password": "wrong"})
+	e.req("GET", "/zones", "jarl_aaaaaaaaaaaaaaaa_bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb", nil)
+	for _, want := range []string{`msg="login failed" ip=127.0.0.1 user="x ip=8.8.8.8"`, `msg="token refused" ip=127.0.0.1 reason=`} {
+		if !strings.Contains(buf.String(), want) {
+			t.Errorf("no %q in\n%s", want, buf.String())
+		}
+	}
+}

internal/server/authn.go

@@ -136,6 +136,8 @@ func (s *Server) authenticate(r *http.Request) (*auth.Principal, *model.Session,
 		p, err := s.tokenPrincipal(r.Context(), strings.TrimSpace(tok), ip, r.UserAgent())
 		if err != nil {
 			s.limiter.fail("ip:" + ip)
+			// For Guard and fail2ban: the address first, so a name can't fake it.
+			s.log.Warn("token refused", "ip", ip, "reason", err.Error())
 			s.eng.Audit(r.Context(), &auth.Principal{IP: ip, Agent: r.UserAgent(), Method: "token"}, "auth.token_failed", "", 0, "", "failure", err.Error())
 			return nil, nil, &httpError{http.StatusUnauthorized, "unauthenticated", "invalid API token: " + err.Error()}
 		}
@@ -328,6 +330,8 @@ func (s *Server) login(w http.ResponseWriter, r *http.Request) error {
 		if err == nil && u.Disabled && ok {
 			reason = "account disabled"
 		}
+		// For Guard and fail2ban: the address first, so a name can't fake it.
+		s.log.Warn("login failed", "ip", ip, "user", name, "reason", reason)
 		s.eng.Audit(r.Context(), &auth.Principal{User: &model.User{Username: name}, IP: ip, Agent: r.UserAgent(), Method: "session"}, "auth.login_failed", "user", 0, name, "failure", reason)
 		return &httpError{http.StatusUnauthorized, "login_failed", "wrong username or password"}
 	}

internal/server/h_acme.go

@@ -198,6 +198,8 @@ func (s *Server) httpreqPrincipal(w http.ResponseWriter, r *http.Request) (*auth
 	}
 	fail := func(reason string) error {
 		s.limiter.fail("ip:" + ip)
+		// For Guard and fail2ban: the address first, so a name can't fake it.
+		s.log.Warn("token refused", "ip", ip, "reason", "httpreq: "+reason)
 		s.eng.Audit(r.Context(), &auth.Principal{User: &model.User{Username: user}, IP: ip, Agent: r.UserAgent(), Method: "token"}, "auth.token_failed", "", 0, "", "failure", "httpreq: "+reason)
 		w.Header().Set("WWW-Authenticate", httpreqRealm)
 		return &httpError{http.StatusUnauthorized, "unauthenticated", "invalid credentials: " + reason}

internal/server/h_totp.go

@@ -206,6 +206,8 @@ func (s *Server) loginTOTP(w http.ResponseWriter, r *http.Request) error {
 	if err != nil {
 		s.limiter.fail("ip:"+ip, userKey)
 		left := s.pendingLogins.fail(in.Challenge)
+		// For Guard and fail2ban: the address first, so a name can't fake it.
+		s.log.Warn("login failed", "ip", ip, "user", u.Username, "reason", "wrong code: "+err.Error())
 		s.eng.Audit(r.Context(), who, "auth.totp_failed", "user", u.ID, u.Username, "failure", err.Error())
 		msg := "wrong code: use the current code from your authenticator app, or a recovery code"
 		if left == 0 {