Decide the ARI replaces retry on the HTTP status, not on the CA's wording

The retry was keyed on the message text and missed ZeroSSL, which answers
HTTP 401 with 'The "replaces" field does not identify a certificate that
belongs to this ACME account', so switching the ACME server failed every
renewal.

RFC 9773 Section 5 mandates a problem type only for the 409
"alreadyReplaced" case and leaves its other checks (same account, shared
identifier) to server policy, so the wording cannot be enumerated. Judge
the status code instead: _isARIReplacesRejected treats only a non-2xx
response as a rejection. That also removes a false positive on the
success path, where the server "MUST reflect that field in the response"
and the echoed base64url certID can itself contain "ARI".

fix #7280
This commit is contained in:
neil
2026-09-26 11:50:22 +02:00 Unverified
parent 879c1519bb
commit 6618d55bfc
+32 -5
View File
@@ -5590,11 +5590,9 @@ issue() {
_on_issue_err "$_post_hook"
return 1
fi
# RFC 9773 Section 5 only defines the "alreadyReplaced" error, but real CAs
# (Let's Encrypt) may also reject with a malformed error if the prior cert
# was issued by a different issuer / different CA. Retry without "replaces"
# whenever the failure mentions ARI or the replaces field.
if [ "$_replaces_certID" ] && { _contains "$response" "alreadyReplaced" || _contains "$response" "urn:ietf:params:acme:error:malformed" || _contains "$response" "'replaces'" || _contains "$response" "ARI"; }; then
# Retry without "replaces" whenever the CA rejected that field, e.g. after
# switching the ACME server: the prior cert belongs to the old CA.
if [ "$_replaces_certID" ] && _isARIReplacesRejected "$code" "$response"; then
_info "ARI 'replaces' rejected by CA, retrying newOrder without 'replaces'."
if ! _send_signed_request "$ACME_NEW_ORDER" "$_newOrderObj}"; then
_err "Error creating new order."
@@ -7930,6 +7928,35 @@ _getARICertID() {
printf "%s.%s" "$_akiurl" "$_serurl"
}
#httpcode response
#Returns 0 when a newOrder was rejected because of the ARI "replaces" field,
#so that the order can be retried without it.
#The status code decides first, and an empty code counts as "not rejected":
#an ACCEPTED order echoes the field back, since RFC 9773 Section 5 says that
#a server accepting a newOrder request with a "replaces" field "MUST reflect
#that field in the response", and the certID it carries is base64url, so the
#response of a SUCCESSFUL order can contain "replaces" and even "ARI".
#Matching on the message alone would then re-order without "replaces" and
#defeat ARI.
#Only the 409 "alreadyReplaced" type is mandated by RFC 9773 Section 5; the
#other checks it lists (same ACME account, shared identifier) are left to
#server policy, so the wording differs per CA: Let's Encrypt answers
#malformed when the prior cert was issued by a different issuer, ZeroSSL
#answers 401 with 'The "replaces" field does not identify a certificate that
#belongs to this ACME account'.
#https://github.com/acmesh-official/acme.sh/issues/7280
_isARIReplacesRejected() {
_ari_rej_code="$1"
_ari_rej_resp="$2"
if [ -z "$_ari_rej_code" ] || _startswith "$_ari_rej_code" "2"; then
return 1
fi
_contains "$_ari_rej_resp" "alreadyReplaced" ||
_contains "$_ari_rej_resp" "replaces" ||
_contains "$_ari_rej_resp" "ARI" ||
_contains "$_ari_rej_resp" "urn:ietf:params:acme:error:malformed"
}
#cert
_get_ARI() {
_cert="$1"