istio.io/index.md at 4a114797f9ca8ce0c5a6b01f520866a4e5aa7d33

18 KiB

Raw Blame History

title

description

weight

keywords

aliases

Basic Authentication Policy

Shows you how to use Istio authentication policy to setup mutual TLS and basic end-user authentication.

security

authentication

/docs/tasks/security/istio-auth.html

Through this task, you will learn how to:

Use authentication policy to setup mutual TLS.
Use authentication policy to do end-user authentication.

Before you begin

Understand Istio authentication policy and related mutual TLS authentication concepts.
Have a Kubernetes cluster with Istio installed, without global mutual TLS enabled (e.g use install/kubernetes/istio-demo.yaml as described in installation steps, or set global.mtls.enabled to false using Helm).

For demo, create two namespaces foo and bar, and deploy httpbin and sleep with sidecar on both of them. Also, run another httpbin and sleep app without sidecar (to keep it separate, run them in legacy namespace). In a regular system, a service can be both server (receiving traffic) for some services, and client for some others. For simplicity, in this demo, we only use sleep apps as clients, and httpbin as servers.

$ kubectl create ns foo
$ kubectl apply -f <(istioctl kube-inject -f @samples/httpbin/httpbin.yaml@) -n foo
$ kubectl apply -f <(istioctl kube-inject -f @samples/sleep/sleep.yaml@) -n foo
$ kubectl create ns bar
$ kubectl apply -f <(istioctl kube-inject -f @samples/httpbin/httpbin.yaml@) -n bar
$ kubectl apply -f <(istioctl kube-inject -f @samples/sleep/sleep.yaml@) -n bar
$ kubectl create ns legacy
$ kubectl apply -f @samples/httpbin/httpbin.yaml@ -n legacy
$ kubectl apply -f @samples/sleep/sleep.yaml@ -n legacy

Verifying setup by sending an HTTP request (using curl command) from any clients (i.e sleep.foo, sleep.bar and sleep.legacy) to any server (httpbin.foo, httpbin.bar or httpbin.legacy). All requests should success with HTTP code 200.

For example, here is a command to check sleep.bar to httpbin.foo reachability:

$ kubectl exec $(kubectl get pod -l app=sleep -n bar -o jsonpath={.items..metadata.name}) -c sleep -n bar -- curl http://httpbin.foo:8000/ip -s -o /dev/null -w "%{http_code}\n"
200

Conveniently, this one-liner command iterates through all combinations:

$ for from in "foo" "bar" "legacy"; do for to in "foo" "bar" "legacy"; do kubectl exec $(kubectl get pod -l app=sleep -n ${from} -o jsonpath={.items..metadata.name}) -c sleep -n ${from} -- curl http://httpbin.${to}:8000/ip -s -o /dev/null -w "sleep.${from} to httpbin.${to}: %{http_code}\n"; done; done
sleep.foo to httpbin.foo: 200
sleep.foo to httpbin.bar: 200
sleep.foo to httpbin.legacy: 200
sleep.bar to httpbin.foo: 200
sleep.bar to httpbin.bar: 200
sleep.bar to httpbin.legacy: 200
sleep.legacy to httpbin.foo: 200
sleep.legacy to httpbin.bar: 200
sleep.legacy to httpbin.legacy: 200

If you have curl installed on istio-proxy container, you can also verify that it can reach httpbin services in from proxy:

$ kubectl exec $(kubectl get pod -l app=sleep -n bar -o jsonpath={.items..metadata.name}) -c istio-proxy -n bar -- curl http://httpbin.foo:8000/ip -s -o /dev/null -w "%{http_code}\n"
200

Last but not least, verify that there are no authentication policy:
```
$ kubectl get meshpolicies.authentication.istio.io
No resources found.
$ kubectl get policies.authentication.istio.io --all-namespaces
No resources found.
$ kubectl get destinationrules.networking.istio.io --all-namespaces
```
You may see some policies and/or destination rules added by Istio installation, depends on installation mode. However, there should be none for foo, bar and legacy namespaces.

Enable mutual TLS for all services in the mesh

To enable mutual TLS for all services in the mesh, you can submit mesh authentication policy and destination rule as below:

cat <<EOF | istioctl create -f -
apiVersion: "authentication.istio.io/v1alpha1"
kind: "MeshPolicy"
metadata:
  name: "default"
spec:
  peers:
  - mtls: {}
EOF

cat <<EOF | istioctl create -f -
apiVersion: "networking.istio.io/v1alpha3"
kind: "DestinationRule"
metadata:
  name: "default"
spec:
  host: "*.local"
  trafficPolicy:
    tls:
      mode: ISTIO_MUTUAL
EOF

Mesh-wide authentication policy name must be default; any policy with other name will be rejected and ignored. Also note that the CRD kind is MeshPolicy, which is different than the namespace-wide or service-specific policy kind (Policy)
On the other hand, the destination rule can have any name, and in any namespace. For consistency, we use also name it default and keep in default namespace in this demo.
Host value *.local in destination rule matches only services in the mesh, which have local suffix.
With ISTIO_MUTUAL TLS mode, Istio will set the path for key and certificates (e.g clientCertificate, privateKey and caCertificates) according to its internal implementation.
If you want to define a destination rule for a specific service, the TLS settings must be copied over to the new rule.

These authentication policy and destination rule effectively configures sidecars of all services to receive and send request in mutual TLS mode, respectively. However, it cannot be applied on services that don't have sidecar, i.e httpbin.legacy and sleep.legacy in this setup. If you run the same testing command as above, you should see requests from sleep.legacy to httpbin.foo and httpbin.bar start to fail, as the result of enabling mutual TLS on server, but sleep.legacy doesn't have a sidecar to support it. Similarly, requests from sleep.foo (or sleep.bar) to httpbin.legacy also fail.

$ for from in "foo" "bar" "legacy"; do for to in "foo" "bar" "legacy"; do kubectl exec $(kubectl get pod -l app=sleep -n ${from} -o jsonpath={.items..metadata.name}) -c sleep -n ${from} -- curl http://httpbin.${to}:8000/ip -s -o /dev/null -w "sleep.${from} to httpbin.${to}: %{http_code}\n"; done; done
sleep.foo to httpbin.foo: 200
sleep.foo to httpbin.bar: 200
sleep.foo to httpbin.legacy: 503
sleep.bar to httpbin.foo: 200
sleep.bar to httpbin.bar: 200
sleep.bar to httpbin.legacy: 503
sleep.legacy to httpbin.foo: 000
command terminated with exit code 56
sleep.legacy to httpbin.bar: 000
command terminated with exit code 56
sleep.legacy to httpbin.legacy: 200

The HTTP return code is not yet consistent. If request is sent in mutual TLS mode and server accept HTTP only, error code is 503. On the contrary, if request is sent in plain-text to server that use mutual TLS, the error code is 000 (with curl exit code 56, "failure with receiving network data").

To fix the connection from client-with-sidecar to server-without-sidecar, you can add destination rule specific for those server to overwrite TLS setting.

cat <<EOF | istioctl create -f -
apiVersion: "networking.istio.io/v1alpha3"
kind: "DestinationRule"
metadata:
  name: "httpbin"
  namespace: "legacy"
spec:
  host: "httpbin.legacy.svc.cluster.local"
  trafficPolicy:
    tls:
      mode: DISABLE
EOF

Retry sending requests to httpbin.legacy, all should work

$ for from in "foo" "bar" "legacy"; do for to in "legacy"; do kubectl exec $(kubectl get pod -l app=sleep -n ${from} -o jsonpath={.items..metadata.name}) -c sleep -n ${from} -- curl http://httpbin.${to}:8000/ip -s -o /dev/null -w "sleep.${from} to httpbin.${to}: %{http_code}\n"; done; done
sleep.foo to httpbin.legacy: 200
sleep.bar to httpbin.legacy: 200
sleep.legacy to httpbin.legacy: 200

This approach can also be used to configure Kubernetes' API server, when mutual TLS is enabled globally. Following is an example configuration.

cat <<EOF | istioctl create -f -
apiVersion: "networking.istio.io/v1alpha3"
kind: "DestinationRule"
metadata:
  name: "api-server"
  namespace: "default"
spec:
  host: "kubernetes.default.svc.cluster.local"
  trafficPolicy:
    tls:
      mode: DISABLE
EOF

For the second issue, connection from client-without-sidecar to server-with-sidecar (in mutual TLS mode), the only option is to drop mutual TLS to PERMISSIVE mode, which allows server to accept traffic in either HTTP or (mutual) TLS. This, obviously, will reduce security level, and is recommended to use during migration only. To do so, you can change the mesh policy (adding mode: PERMISSIVE under mtls block). A more conservative (and recommended) way is creating new policy only for the specific service(s) needed. The example below illustrates the latter:

cat <<EOF | istioctl create -f -
apiVersion: "authentication.istio.io/v1alpha1"
kind: "Policy"
metadata:
  name: "httpbin"
  namespace: "foo"
spec:
  targets:
  - name: "httpbin"
  peers:
  - mtls:
      mode: PERMISSIVE
EOF

Request from sleep.legacy to httpbin.foo should succeed, while to httpbin.bar still fail.

$ kubectl exec $(kubectl get pod -l app=sleep -n legacy -o jsonpath={.items..metadata.name}) -c sleep -n legacy -- curl http://httpbin.foo:8000/ip -s -o /dev/null -w "%{http_code}\n"
200
$ kubectl exec $(kubectl get pod -l app=sleep -n legacy -o jsonpath={.items..metadata.name}) -c sleep -n legacy -- curl http://httpbin.bar:8000/ip -s -o /dev/null -w "%{http_code}\n"
000

Before move on the next section, let's remove authentication policies and destination rules created in this section

kubectl delete meshpolicy.authentication.istio.io default
kubectl delete policy.authentication.istio.io -n foo --all
kubectl delete destinationrules.networking.istio.io default
kubectl delete destinationrules.networking.istio.io -n legacy --all

Enable mutual TLS for all services in a namespace

Instead of enabling mutual TLS globally, you can do it per namespace. The process is similar, except the policy is in namespace-scope (kind Policy)

cat <<EOF | istioctl create -f -
apiVersion: "authentication.istio.io/v1alpha1"
kind: "Policy"
metadata:
  name: "default"
  namespace: "foo"
spec:
  peers:
  - mtls: {}
EOF

Similar to mesh-wide policy, namespace-wide policy must be named default, and doesn't restrict any specific service (no targets section)

Add corresponding destination rule:

cat <<EOF | istioctl create -f -
apiVersion: "networking.istio.io/v1alpha3"
kind: "DestinationRule"
metadata:
  name: "default"
  namespace: "foo"
spec:
  host: "*.foo.svc.cluster.local"
  trafficPolicy:
    tls:
      mode: ISTIO_MUTUAL
EOF

Host *.foo.svc.cluster.local limits the matches to services in foo namespace only.

As these policy and destination rule are applied on services in namespace foo only, you should see only request from client-without-sidecar (sleep.legacy) to httpbin.foo start to fail.

$ for from in "foo" "bar" "legacy"; do for to in "foo" "bar" "legacy"; do kubectl exec $(kubectl get pod -l app=sleep -n ${from} -o jsonpath={.items..metadata.name}) -c sleep -n ${from} -- curl http://httpbin.${to}:8000/ip -s -o /dev/null -w "sleep.${from} to httpbin.${to}: %{http_code}\n"; done; done
sleep.foo to httpbin.foo: 200
sleep.foo to httpbin.bar: 200
sleep.foo to httpbin.legacy: 200
sleep.bar to httpbin.foo: 200
sleep.bar to httpbin.bar: 200
sleep.bar to httpbin.legacy: 200
sleep.legacy to httpbin.foo: 000
command terminated with exit code 56
sleep.legacy to httpbin.bar: 200
sleep.legacy to httpbin.legacy: 200

Enable mutual TLS for single service `httpbin.bar`

You can also set authentication policy and destination rule for a specific service. Run this command to set another policy only for httpbin.bar service.

cat <<EOF | istioctl create -n bar -f -
apiVersion: "authentication.istio.io/v1alpha1"
kind: "Policy"
metadata:
  name: "httpbin"
spec:
  targets:
  - name: httpbin
  peers:
  - mtls: {}
EOF

And a destination rule:

cat <<EOF | istioctl create -n bar -f -
apiVersion: "networking.istio.io/v1alpha3"
kind: "DestinationRule"
metadata:
  name: "httpbin"
spec:
  host: "httpbin.bar.svc.cluster.local"
  trafficPolicy:
    tls:
      mode: ISTIO_MUTUAL
EOF

In this example, we do not specify namespace in metadata but put it in the command line (-n bar). They should work the same. There is no restriction on the authentication policy and destination rule name. The example use the name of the service itself for simplicity.

Again, run the probing command. As expected, request from sleep.legacy to httpbin.bar starts failing with the same reasons.

...
sleep.legacy to httpbin.bar: 000
command terminated with exit code 56

If we have more services in namespace bar, we should see traffic to them won't be affected. Instead of adding more services to demonstrate this behavior, we edit the policy slightly to apply on a specific port:

cat <<EOF | istioctl replace -n bar -f -
apiVersion: "authentication.istio.io/v1alpha1"
kind: "Policy"
metadata:
  name: "example-2"
spec:
  targets:
  - name: httpbin
    ports:
    - number: 1234
  peers:
  - mtls:
EOF

And a corresponding change to the destination rule:

cat <<EOF | istioctl replace -n bar -f -
apiVersion: "networking.istio.io/v1alpha3"
kind: "DestinationRule"
metadata:
  name: "example-2"
spec:
  host: httpbin.bar.svc.cluster.local
  trafficPolicy:
    tls:
      mode: DISABLE
    portLevelSettings:
    - port:
        number: 1234
      tls:
        mode: ISTIO_MUTUAL
EOF

This new policy will apply only to the httpbin service on port 1234. As a result, mutual TLS is disabled (again) on port 8000 and requests from sleep.legacy will resume working.

$ kubectl exec $(kubectl get pod -l app=sleep -n legacy -o jsonpath={.items..metadata.name}) -c sleep -n legacy -- curl http://httpbin.bar:8000/ip -s -o /dev/null -w "%{http_code}\n"
200

Having both namespace-level and service-level policies

Assuming we already added the namespace-level policy that enables mutual TLS for all services in namespace foo and observe that request from sleep.legacy to httpbin.foo are failing (see above). Now add another policy that disables mutual TLS (peers section is empty) specifically for the httpbin service:

cat <<EOF | istioctl create -n foo -f -
apiVersion: "authentication.istio.io/v1alpha1"
kind: "Policy"
metadata:
  name: "example-3"
spec:
  targets:
  - name: httpbin
EOF

and destination rule:

cat <<EOF | istioctl create -n foo -f -
apiVersion: "networking.istio.io/v1alpha3"
kind: "DestinationRule"
metadata:
  name: "example-3"
spec:
  host: httpbin.foo.svc.cluster.local
  trafficPolicy:
    tls:
      mode: DISABLE
EOF

Re-run the request from sleep.legacy, we should see a success return code again (200), confirming service-level policy overrules the namespace-level policy.

$ kubectl exec $(kubectl get pod -l app=sleep -n legacy -o jsonpath={.items..metadata.name}) -c sleep -n legacy -- curl http://httpbin.foo:8000/ip -s -o /dev/null -w "%{http_code}\n"
200

Setup end-user authentication

You will need a valid JWT (corresponding to the JWKS endpoint you want to use for the demo). Please follow the instructions here to create one. You can also use your own JWT/JWKS endpoint for the demo. Once you have that, export to some environment variables.

$ export SVC_ACCOUNT="example@my-project.iam.gserviceaccount.com"
$ export JWKS=https://www.googleapis.com/service_accounts/v1/jwk/${SVC_ACCOUNT}
$ export TOKEN=<YOUR-TOKEN>

Also, for convenience, expose httpbin.foo via ingress (for more details, see ingress task).

cat <<EOF | kubectl apply -f -
apiVersion: extensions/v1beta1
kind: Ingress
metadata:
  name: httpbin-ingress
  namespace: foo
  annotations:
    kubernetes.io/ingress.class: istio
spec:
  rules:
  - http:
      paths:
      - path: /headers
        backend:
          serviceName: httpbin
          servicePort: 8000
EOF

Get ingress IP

export INGRESS_HOST=$(kubectl get ing -n foo -o=jsonpath='{.items[0].status.loadBalancer.ingress[0].ip}')

And run a test query

$ curl $INGRESS_HOST/headers -s -o /dev/null -w "%{http_code}\n"
200

Now, let's add a policy that requires end-user JWT for httpbin.foo. The next command assumes policy with name "httpbin" already exists (which should be if you follow previous sections). You can run kubectl get policies.authentication.istio.io -n foo to confirm, and use istio create (instead of istio replace) if resource is not found. Also note in this policy, peer authentication (mutual TLS) is also set, though it can be removed without affecting origin authentication settings.

cat <<EOF | istioctl replace -n foo -f -
apiVersion: "authentication.istio.io/v1alpha1"
kind: "Policy"
metadata:
  name: "example-3"
spec:
  targets:
  - name: httpbin
  peers:
  - mtls:
  origins:
  - jwt:
      issuer: $SVC_ACCOUNT
      jwksUri: $JWKS
  principalBinding: USE_ORIGIN
EOF

The same curl command from before will return with 401 error code, as a result of server is expecting JWT but none was provided:

$ curl $INGRESS_HOST/headers -s -o /dev/null -w "%{http_code}\n"
401

Attaching the valid token generated above returns success:

$ curl --header "Authorization: Bearer $TOKEN" $INGRESS_HOST/headers -s -o /dev/null -w "%{http_code}\n"
200

You may want to try to modify token or policy (e.g change issuer, audiences, expiry date etc) to observe other aspects of JWT validation.

Cleanup

Remove all resources.

$ kubectl delete ns foo bar legacy

What's next

Learn more about verifying mutual TLS setup testing Istio mutual TLS authentication

18 KiB Raw Blame History