certmagic/httphandlers.go

211 lines
7.5 KiB
Go
Raw Permalink Normal View History

2018-12-10 13:15:26 +10:00
// Copyright 2015 Matthew Holt
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package certmagic
import (
"net/http"
"net/url"
2018-12-10 13:15:26 +10:00
"strings"
"github.com/mholt/acmez/v2/acme"
"go.uber.org/zap"
2018-12-10 13:15:26 +10:00
)
// HTTPChallengeHandler wraps h in a handler that can solve the ACME
// HTTP challenge. cfg is required, and it must have a certificate
// cache backed by a functional storage facility, since that is where
// the challenge state is stored between initiation and solution.
//
// If a request is not an ACME HTTP challenge, h will be invoked.
func (am *ACMEIssuer) HTTPChallengeHandler(h http.Handler) http.Handler {
2018-12-10 13:15:26 +10:00
return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
Major refactor to improve performance, correctness, and extensibility Breaking changes; thank goodness we're not 1.0 yet 😅 - read on! This change completely separates ACME-specific code from the rest of the certificate management process, allowing pluggable sources for certs that aren't ACME. Notably, most of Config was spliced into ACMEManager. Similarly, there's now Default and DefaultACME. Storage structure had to be reconfigured. Certificates are no longer in the acme/ subfolder since they can be obtained by ways other than ACME! Certificates moved to a new certificates/ subfolder. The subfolders in that folder use the path of the ACME endpoint instead of just the host, so that also changed. Be aware that unless you move your certs over, CertMagic will not find them and will attempt to get new ones. That is usually fine for most users, but for extremely large deployments, you will want to move them over first. Old certs path: acme/acme-staging-v02.api.letsencrypt.org/... New certs path: certificates/acme-staging-v02.api.letsencrypt.org-directory/... That's all for significant storage changes! But this refactor also vastly improves performance, especially at scale, and makes CertMagic way more resilient to errors. Retries are done on the staging endpoint by default, so they won't count against your rate limit. If your hardware can handle it, I'm now pretty confident that you can give CertMagic a million domain names and it will gracefully manage them, as fast as it can within internal and external rate limits, even in the presence of errors. Errors will of course slow some things down, but you should be good to go if you're monitoring logs and can fix any misconfigurations or other external errors! Several other mostly-minor enhancements fix bugs, especially at scale. For example, duplicated renewal tasks (that continuously fail) will not pile up on each other: only one will operate, under exponential backoff. Closes #50 and fixes #55
2020-02-22 07:32:57 +10:00
if am.HandleHTTPChallenge(w, r) {
2018-12-10 13:15:26 +10:00
return
}
h.ServeHTTP(w, r)
})
}
Major refactor to improve performance, correctness, and extensibility Breaking changes; thank goodness we're not 1.0 yet 😅 - read on! This change completely separates ACME-specific code from the rest of the certificate management process, allowing pluggable sources for certs that aren't ACME. Notably, most of Config was spliced into ACMEManager. Similarly, there's now Default and DefaultACME. Storage structure had to be reconfigured. Certificates are no longer in the acme/ subfolder since they can be obtained by ways other than ACME! Certificates moved to a new certificates/ subfolder. The subfolders in that folder use the path of the ACME endpoint instead of just the host, so that also changed. Be aware that unless you move your certs over, CertMagic will not find them and will attempt to get new ones. That is usually fine for most users, but for extremely large deployments, you will want to move them over first. Old certs path: acme/acme-staging-v02.api.letsencrypt.org/... New certs path: certificates/acme-staging-v02.api.letsencrypt.org-directory/... That's all for significant storage changes! But this refactor also vastly improves performance, especially at scale, and makes CertMagic way more resilient to errors. Retries are done on the staging endpoint by default, so they won't count against your rate limit. If your hardware can handle it, I'm now pretty confident that you can give CertMagic a million domain names and it will gracefully manage them, as fast as it can within internal and external rate limits, even in the presence of errors. Errors will of course slow some things down, but you should be good to go if you're monitoring logs and can fix any misconfigurations or other external errors! Several other mostly-minor enhancements fix bugs, especially at scale. For example, duplicated renewal tasks (that continuously fail) will not pile up on each other: only one will operate, under exponential backoff. Closes #50 and fixes #55
2020-02-22 07:32:57 +10:00
// HandleHTTPChallenge uses am to solve challenge requests from an ACME
2018-12-10 13:15:26 +10:00
// server that were initiated by this instance or any other instance in
Major refactor to improve performance, correctness, and extensibility Breaking changes; thank goodness we're not 1.0 yet 😅 - read on! This change completely separates ACME-specific code from the rest of the certificate management process, allowing pluggable sources for certs that aren't ACME. Notably, most of Config was spliced into ACMEManager. Similarly, there's now Default and DefaultACME. Storage structure had to be reconfigured. Certificates are no longer in the acme/ subfolder since they can be obtained by ways other than ACME! Certificates moved to a new certificates/ subfolder. The subfolders in that folder use the path of the ACME endpoint instead of just the host, so that also changed. Be aware that unless you move your certs over, CertMagic will not find them and will attempt to get new ones. That is usually fine for most users, but for extremely large deployments, you will want to move them over first. Old certs path: acme/acme-staging-v02.api.letsencrypt.org/... New certs path: certificates/acme-staging-v02.api.letsencrypt.org-directory/... That's all for significant storage changes! But this refactor also vastly improves performance, especially at scale, and makes CertMagic way more resilient to errors. Retries are done on the staging endpoint by default, so they won't count against your rate limit. If your hardware can handle it, I'm now pretty confident that you can give CertMagic a million domain names and it will gracefully manage them, as fast as it can within internal and external rate limits, even in the presence of errors. Errors will of course slow some things down, but you should be good to go if you're monitoring logs and can fix any misconfigurations or other external errors! Several other mostly-minor enhancements fix bugs, especially at scale. For example, duplicated renewal tasks (that continuously fail) will not pile up on each other: only one will operate, under exponential backoff. Closes #50 and fixes #55
2020-02-22 07:32:57 +10:00
// this cluster (being, any instances using the same storage am does).
2018-12-10 13:15:26 +10:00
//
// If the HTTP challenge is disabled, this function is a no-op.
//
Major refactor to improve performance, correctness, and extensibility Breaking changes; thank goodness we're not 1.0 yet 😅 - read on! This change completely separates ACME-specific code from the rest of the certificate management process, allowing pluggable sources for certs that aren't ACME. Notably, most of Config was spliced into ACMEManager. Similarly, there's now Default and DefaultACME. Storage structure had to be reconfigured. Certificates are no longer in the acme/ subfolder since they can be obtained by ways other than ACME! Certificates moved to a new certificates/ subfolder. The subfolders in that folder use the path of the ACME endpoint instead of just the host, so that also changed. Be aware that unless you move your certs over, CertMagic will not find them and will attempt to get new ones. That is usually fine for most users, but for extremely large deployments, you will want to move them over first. Old certs path: acme/acme-staging-v02.api.letsencrypt.org/... New certs path: certificates/acme-staging-v02.api.letsencrypt.org-directory/... That's all for significant storage changes! But this refactor also vastly improves performance, especially at scale, and makes CertMagic way more resilient to errors. Retries are done on the staging endpoint by default, so they won't count against your rate limit. If your hardware can handle it, I'm now pretty confident that you can give CertMagic a million domain names and it will gracefully manage them, as fast as it can within internal and external rate limits, even in the presence of errors. Errors will of course slow some things down, but you should be good to go if you're monitoring logs and can fix any misconfigurations or other external errors! Several other mostly-minor enhancements fix bugs, especially at scale. For example, duplicated renewal tasks (that continuously fail) will not pile up on each other: only one will operate, under exponential backoff. Closes #50 and fixes #55
2020-02-22 07:32:57 +10:00
// If am is nil or if am does not have a certificate cache backed by
2018-12-10 13:15:26 +10:00
// usable storage, solving the HTTP challenge will fail.
//
// It returns true if it handled the request; if so, the response has
// already been written. If false is returned, this call was a no-op and
// the request has not been handled.
func (am *ACMEIssuer) HandleHTTPChallenge(w http.ResponseWriter, r *http.Request) bool {
Major refactor to improve performance, correctness, and extensibility Breaking changes; thank goodness we're not 1.0 yet 😅 - read on! This change completely separates ACME-specific code from the rest of the certificate management process, allowing pluggable sources for certs that aren't ACME. Notably, most of Config was spliced into ACMEManager. Similarly, there's now Default and DefaultACME. Storage structure had to be reconfigured. Certificates are no longer in the acme/ subfolder since they can be obtained by ways other than ACME! Certificates moved to a new certificates/ subfolder. The subfolders in that folder use the path of the ACME endpoint instead of just the host, so that also changed. Be aware that unless you move your certs over, CertMagic will not find them and will attempt to get new ones. That is usually fine for most users, but for extremely large deployments, you will want to move them over first. Old certs path: acme/acme-staging-v02.api.letsencrypt.org/... New certs path: certificates/acme-staging-v02.api.letsencrypt.org-directory/... That's all for significant storage changes! But this refactor also vastly improves performance, especially at scale, and makes CertMagic way more resilient to errors. Retries are done on the staging endpoint by default, so they won't count against your rate limit. If your hardware can handle it, I'm now pretty confident that you can give CertMagic a million domain names and it will gracefully manage them, as fast as it can within internal and external rate limits, even in the presence of errors. Errors will of course slow some things down, but you should be good to go if you're monitoring logs and can fix any misconfigurations or other external errors! Several other mostly-minor enhancements fix bugs, especially at scale. For example, duplicated renewal tasks (that continuously fail) will not pile up on each other: only one will operate, under exponential backoff. Closes #50 and fixes #55
2020-02-22 07:32:57 +10:00
if am == nil {
2018-12-10 13:15:26 +10:00
return false
}
Major refactor to improve performance, correctness, and extensibility Breaking changes; thank goodness we're not 1.0 yet 😅 - read on! This change completely separates ACME-specific code from the rest of the certificate management process, allowing pluggable sources for certs that aren't ACME. Notably, most of Config was spliced into ACMEManager. Similarly, there's now Default and DefaultACME. Storage structure had to be reconfigured. Certificates are no longer in the acme/ subfolder since they can be obtained by ways other than ACME! Certificates moved to a new certificates/ subfolder. The subfolders in that folder use the path of the ACME endpoint instead of just the host, so that also changed. Be aware that unless you move your certs over, CertMagic will not find them and will attempt to get new ones. That is usually fine for most users, but for extremely large deployments, you will want to move them over first. Old certs path: acme/acme-staging-v02.api.letsencrypt.org/... New certs path: certificates/acme-staging-v02.api.letsencrypt.org-directory/... That's all for significant storage changes! But this refactor also vastly improves performance, especially at scale, and makes CertMagic way more resilient to errors. Retries are done on the staging endpoint by default, so they won't count against your rate limit. If your hardware can handle it, I'm now pretty confident that you can give CertMagic a million domain names and it will gracefully manage them, as fast as it can within internal and external rate limits, even in the presence of errors. Errors will of course slow some things down, but you should be good to go if you're monitoring logs and can fix any misconfigurations or other external errors! Several other mostly-minor enhancements fix bugs, especially at scale. For example, duplicated renewal tasks (that continuously fail) will not pile up on each other: only one will operate, under exponential backoff. Closes #50 and fixes #55
2020-02-22 07:32:57 +10:00
if am.DisableHTTPChallenge {
2018-12-10 13:15:26 +10:00
return false
}
if !LooksLikeHTTPChallenge(r) {
2018-12-10 13:15:26 +10:00
return false
}
Major refactor to improve performance, correctness, and extensibility Breaking changes; thank goodness we're not 1.0 yet 😅 - read on! This change completely separates ACME-specific code from the rest of the certificate management process, allowing pluggable sources for certs that aren't ACME. Notably, most of Config was spliced into ACMEManager. Similarly, there's now Default and DefaultACME. Storage structure had to be reconfigured. Certificates are no longer in the acme/ subfolder since they can be obtained by ways other than ACME! Certificates moved to a new certificates/ subfolder. The subfolders in that folder use the path of the ACME endpoint instead of just the host, so that also changed. Be aware that unless you move your certs over, CertMagic will not find them and will attempt to get new ones. That is usually fine for most users, but for extremely large deployments, you will want to move them over first. Old certs path: acme/acme-staging-v02.api.letsencrypt.org/... New certs path: certificates/acme-staging-v02.api.letsencrypt.org-directory/... That's all for significant storage changes! But this refactor also vastly improves performance, especially at scale, and makes CertMagic way more resilient to errors. Retries are done on the staging endpoint by default, so they won't count against your rate limit. If your hardware can handle it, I'm now pretty confident that you can give CertMagic a million domain names and it will gracefully manage them, as fast as it can within internal and external rate limits, even in the presence of errors. Errors will of course slow some things down, but you should be good to go if you're monitoring logs and can fix any misconfigurations or other external errors! Several other mostly-minor enhancements fix bugs, especially at scale. For example, duplicated renewal tasks (that continuously fail) will not pile up on each other: only one will operate, under exponential backoff. Closes #50 and fixes #55
2020-02-22 07:32:57 +10:00
return am.distributedHTTPChallengeSolver(w, r)
2018-12-10 13:15:26 +10:00
}
// distributedHTTPChallengeSolver checks to see if this challenge
// request was initiated by this or another instance which uses the
Major refactor to improve performance, correctness, and extensibility Breaking changes; thank goodness we're not 1.0 yet 😅 - read on! This change completely separates ACME-specific code from the rest of the certificate management process, allowing pluggable sources for certs that aren't ACME. Notably, most of Config was spliced into ACMEManager. Similarly, there's now Default and DefaultACME. Storage structure had to be reconfigured. Certificates are no longer in the acme/ subfolder since they can be obtained by ways other than ACME! Certificates moved to a new certificates/ subfolder. The subfolders in that folder use the path of the ACME endpoint instead of just the host, so that also changed. Be aware that unless you move your certs over, CertMagic will not find them and will attempt to get new ones. That is usually fine for most users, but for extremely large deployments, you will want to move them over first. Old certs path: acme/acme-staging-v02.api.letsencrypt.org/... New certs path: certificates/acme-staging-v02.api.letsencrypt.org-directory/... That's all for significant storage changes! But this refactor also vastly improves performance, especially at scale, and makes CertMagic way more resilient to errors. Retries are done on the staging endpoint by default, so they won't count against your rate limit. If your hardware can handle it, I'm now pretty confident that you can give CertMagic a million domain names and it will gracefully manage them, as fast as it can within internal and external rate limits, even in the presence of errors. Errors will of course slow some things down, but you should be good to go if you're monitoring logs and can fix any misconfigurations or other external errors! Several other mostly-minor enhancements fix bugs, especially at scale. For example, duplicated renewal tasks (that continuously fail) will not pile up on each other: only one will operate, under exponential backoff. Closes #50 and fixes #55
2020-02-22 07:32:57 +10:00
// same storage as am does, and attempts to complete the challenge for
2018-12-10 13:15:26 +10:00
// it. It returns true if the request was handled; false otherwise.
func (am *ACMEIssuer) distributedHTTPChallengeSolver(w http.ResponseWriter, r *http.Request) bool {
Major refactor to improve performance, correctness, and extensibility Breaking changes; thank goodness we're not 1.0 yet 😅 - read on! This change completely separates ACME-specific code from the rest of the certificate management process, allowing pluggable sources for certs that aren't ACME. Notably, most of Config was spliced into ACMEManager. Similarly, there's now Default and DefaultACME. Storage structure had to be reconfigured. Certificates are no longer in the acme/ subfolder since they can be obtained by ways other than ACME! Certificates moved to a new certificates/ subfolder. The subfolders in that folder use the path of the ACME endpoint instead of just the host, so that also changed. Be aware that unless you move your certs over, CertMagic will not find them and will attempt to get new ones. That is usually fine for most users, but for extremely large deployments, you will want to move them over first. Old certs path: acme/acme-staging-v02.api.letsencrypt.org/... New certs path: certificates/acme-staging-v02.api.letsencrypt.org-directory/... That's all for significant storage changes! But this refactor also vastly improves performance, especially at scale, and makes CertMagic way more resilient to errors. Retries are done on the staging endpoint by default, so they won't count against your rate limit. If your hardware can handle it, I'm now pretty confident that you can give CertMagic a million domain names and it will gracefully manage them, as fast as it can within internal and external rate limits, even in the presence of errors. Errors will of course slow some things down, but you should be good to go if you're monitoring logs and can fix any misconfigurations or other external errors! Several other mostly-minor enhancements fix bugs, especially at scale. For example, duplicated renewal tasks (that continuously fail) will not pile up on each other: only one will operate, under exponential backoff. Closes #50 and fixes #55
2020-02-22 07:32:57 +10:00
if am == nil {
2018-12-10 13:15:26 +10:00
return false
}
host := hostOnly(r.Host)
chalInfo, distributed, err := am.config.getChallengeInfo(r.Context(), host)
2018-12-10 13:15:26 +10:00
if err != nil {
am.Logger.Warn("looking up info for HTTP challenge",
zap.String("host", host),
zap.String("remote_addr", r.RemoteAddr),
zap.String("user_agent", r.Header.Get("User-Agent")),
zap.Error(err))
2018-12-10 13:15:26 +10:00
return false
}
return solveHTTPChallenge(am.Logger, w, r, chalInfo.Challenge, distributed)
2018-12-10 13:15:26 +10:00
}
// solveHTTPChallenge solves the HTTP challenge using the given challenge information.
// If the challenge is being solved in a distributed fahsion, set distributed to true for logging purposes.
// It returns true the properties of the request check out in relation to the HTTP challenge.
// Most of this code borrowed from xenolf's built-in HTTP-01 challenge solver in March 2018.
func solveHTTPChallenge(logger *zap.Logger, w http.ResponseWriter, r *http.Request, challenge acme.Challenge, distributed bool) bool {
challengeReqPath := challenge.HTTP01ResourcePath()
2018-12-10 13:15:26 +10:00
if r.URL.Path == challengeReqPath &&
strings.EqualFold(hostOnly(r.Host), challenge.Identifier.Value) && // mitigate DNS rebinding attacks
r.Method == http.MethodGet {
2018-12-10 13:15:26 +10:00
w.Header().Add("Content-Type", "text/plain")
w.Write([]byte(challenge.KeyAuthorization))
2018-12-10 13:15:26 +10:00
r.Close = true
logger.Info("served key authentication",
zap.String("identifier", challenge.Identifier.Value),
zap.String("challenge", "http-01"),
zap.String("remote", r.RemoteAddr),
zap.Bool("distributed", distributed))
2018-12-10 13:15:26 +10:00
return true
}
return false
}
// SolveHTTPChallenge solves the HTTP challenge. It should be used only on HTTP requests that are
// from ACME servers trying to validate an identifier (i.e. LooksLikeHTTPChallenge() == true). It
// returns true if the request criteria check out and it answered with key authentication, in which
// case no further handling of the request is necessary.
func SolveHTTPChallenge(logger *zap.Logger, w http.ResponseWriter, r *http.Request, challenge acme.Challenge) bool {
return solveHTTPChallenge(logger, w, r, challenge, false)
}
// LooksLikeHTTPChallenge returns true if r looks like an ACME
// HTTP challenge request from an ACME server.
func LooksLikeHTTPChallenge(r *http.Request) bool {
return r.Method == http.MethodGet &&
strings.HasPrefix(r.URL.Path, acmeHTTPChallengeBasePath)
}
// LooksLikeZeroSSLHTTPValidation returns true if the request appears to be
// domain validation from a ZeroSSL/Sectigo CA. NOTE: This API is
// non-standard and is subject to change.
func LooksLikeZeroSSLHTTPValidation(r *http.Request) bool {
return r.Method == http.MethodGet &&
strings.HasPrefix(r.URL.Path, zerosslHTTPValidationBasePath)
}
// HTTPValidationHandler wraps the ZeroSSL HTTP validation handler such that
// it can pass verification checks from ZeroSSL's API.
//
// If a request is not a ZeroSSL HTTP validation request, h will be invoked.
func (iss *ZeroSSLIssuer) HTTPValidationHandler(h http.Handler) http.Handler {
return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
if iss.HandleZeroSSLHTTPValidation(w, r) {
return
}
h.ServeHTTP(w, r)
})
}
// HandleZeroSSLHTTPValidation is to ZeroSSL API HTTP validation requests like HandleHTTPChallenge
// is to ACME HTTP challenge requests.
func (iss *ZeroSSLIssuer) HandleZeroSSLHTTPValidation(w http.ResponseWriter, r *http.Request) bool {
if iss == nil {
return false
}
if !LooksLikeZeroSSLHTTPValidation(r) {
return false
}
return iss.distributedHTTPValidationAnswer(w, r)
}
func (iss *ZeroSSLIssuer) distributedHTTPValidationAnswer(w http.ResponseWriter, r *http.Request) bool {
if iss == nil {
return false
}
logger := iss.Logger
if logger == nil {
logger = zap.NewNop()
}
host := hostOnly(r.Host)
valInfo, distributed, err := iss.getDistributedValidationInfo(r.Context(), host)
if err != nil {
logger.Warn("looking up info for HTTP validation",
zap.String("host", host),
zap.String("remote_addr", r.RemoteAddr),
zap.String("user_agent", r.Header.Get("User-Agent")),
zap.Error(err))
return false
}
return answerHTTPValidation(logger, w, r, valInfo, distributed)
}
func answerHTTPValidation(logger *zap.Logger, rw http.ResponseWriter, req *http.Request, valInfo acme.Challenge, distributed bool) bool {
// ensure URL matches
validationURL, err := url.Parse(valInfo.URL)
if err != nil {
logger.Error("got invalid URL from CA",
zap.String("file_validation_url", valInfo.URL),
zap.Error(err))
rw.WriteHeader(http.StatusInternalServerError)
return true
}
if req.URL.Path != validationURL.Path {
rw.WriteHeader(http.StatusNotFound)
return true
}
rw.Header().Add("Content-Type", "text/plain")
req.Close = true
rw.Write([]byte(valInfo.Token))
logger.Info("served HTTP validation credential",
zap.String("validation_path", valInfo.URL),
zap.String("challenge", "http-01"),
zap.String("remote", req.RemoteAddr),
zap.Bool("distributed", distributed))
return true
}
const (
acmeHTTPChallengeBasePath = "/.well-known/acme-challenge"
zerosslHTTPValidationBasePath = "/.well-known/pki-validation/"
)