README
¶
Nginx Ingress Controller
This is a nginx Ingress controller that uses ConfigMap to store the nginx configuration. See Ingress controller documentation for details on how it works.
Contents
- Conventions
- Requirements
- Dry running
- Deployment
- HTTP
- HTTPS
- TCP Services
- UDP Services
- Proxy Protocol
- NGINX customization
- NGINX status page
- Disabling NGINX ingress controller
- Debug & Troubleshooting
- Limitations
- NGINX Notes
Conventions
Anytime we reference a tls secret, we mean (x509, pem encoded, RSA 2048, etc). You can generate such a certificate with:
openssl req -x509 -nodes -days 365 -newkey rsa:2048 -keyout $(KEY) -out $(CERT) -subj "/CN=$(HOST)/O=$(HOST)"
and creat the secret via kubectl create secret tls --key file --cert file
Requirements
- Default backend 404-server
Dry running the Ingress controller
Before deploying the controller to production you might want to run it outside the cluster and observe it.
$ make controller
$ mkdir /etc/nginx-ssl
$ ./nginx-ingress-controller --running-in-cluster=false --default-backend-service=kube-system/default-http-backend
Deployment
First create a default backend:
$ kubectl create -f examples/default-backend.yaml
$ kubectl expose rc default-http-backend --port=80 --target-port=8080 --name=default-http-backend
Loadbalancers are created via a ReplicationController or Daemonset:
$ kubectl create -f examples/default/rc-default.yaml
HTTP
First we need to deploy some application to publish. To keep this simple we will use the echoheaders app that just returns information about the http request as output
kubectl run echoheaders --image=gcr.io/google_containers/echoserver:1.4 --replicas=1 --port=8080
Now we expose the same application in two different services (so we can create different Ingress rules)
kubectl expose deployment echoheaders --port=80 --target-port=8080 --name=echoheaders-x
kubectl expose deployment echoheaders --port=80 --target-port=8080 --name=echoheaders-y
Next we create a couple of Ingress rules
kubectl create -f examples/ingress.yaml
we check that ingress rules are defined:
$ kubectl get ing
NAME RULE BACKEND ADDRESS
echomap -
foo.bar.com
/foo echoheaders-x:80
bar.baz.com
/bar echoheaders-y:80
/foo echoheaders-x:80
Before the deploy of the Ingress controller we need a default backend 404-server
kubectl create -f examples/default-backend.yaml
kubectl expose rc default-http-backend --port=80 --target-port=8080 --name=default-http-backend
Check NGINX it is running with the defined Ingress rules:
$ LBIP=$(kubectl get node `kubectl get po -l name=nginx-ingress-lb --template '{{range .items}}{{.spec.nodeName}}{{end}}'` --template '{{range $i, $n := .status.addresses}}{{if eq $n.type "ExternalIP"}}{{$n.address}}{{end}}{{end}}')
$ curl $LBIP/foo -H 'Host: foo.bar.com'
HTTPS
You can secure an Ingress by specifying a secret that contains a TLS private key and certificate. Currently the Ingress only supports a single TLS port, 443, and assumes TLS termination. This controller supports SNI. The TLS secret must contain keys named tls.crt and tls.key that contain the certificate and private key to use for TLS, eg:
apiVersion: v1
data:
tls.crt: base64 encoded cert
tls.key: base64 encoded key
kind: Secret
metadata:
name: testsecret
namespace: default
type: Opaque
Referencing this secret in an Ingress will tell the Ingress controller to secure the channel from the client to the loadbalancer using TLS:
apiVersion: extensions/v1beta1
kind: Ingress
metadata:
name: no-rules-map
spec:
tls:
secretName: testsecret
backend:
serviceName: s1
servicePort: 80
Please follow test.sh as a guide on how to generate secrets containing SSL certificates. The name of the secret can be different than the name of the certificate.
Check the example
Default SSL Certificate
NGINX provides the option default_server to allow a catch-all server in case of request with a not configured server name. This configuration works without issues for HTTP traffic.
In case of HTTPS NGINX requires a certificate. For this reason the Ingress controller provides the flag --default-ssl-certificate. The secret behind this flag contains the default certificate to be used in the mentioned case.
If this flag is not provided NGINX will reject the request with the HTTP code 444.
Running without the flag --default-ssl-certificate:
$ curl -v https://10.2.78.7:443
* Rebuilt URL to: https://10.2.78.7:443/
* Trying 10.2.78.7...
* Connected to 10.2.78.7 (10.2.78.7) port 443 (#0)
* ALPN, offering http/1.1
* Cipher selection: ALL:!EXPORT:!EXPORT40:!EXPORT56:!aNULL:!LOW:!RC4:@STRENGTH
* successfully set certificate verify locations:
* CAfile: /etc/ssl/certs/ca-certificates.crt
CApath: /etc/ssl/certs
* TLSv1.2 (OUT), TLS header, Certificate Status (22):
* TLSv1.2 (OUT), TLS handshake, Client hello (1):
* Unknown SSL protocol error in connection to 10.2.78.7:443
* Closing connection 0
curl: (35) Unknown SSL protocol error in connection to 10.2.78.7:443
Specifyng --default-ssl-certificate=default/foo-tls:
core@localhost ~ $ curl -v https://10.2.78.7:443 -k
* Rebuilt URL to: https://10.2.78.7:443/
* Trying 10.2.78.7...
* Connected to 10.2.78.7 (10.2.78.7) port 443 (#0)
* ALPN, offering http/1.1
* Cipher selection: ALL:!EXPORT:!EXPORT40:!EXPORT56:!aNULL:!LOW:!RC4:@STRENGTH
* successfully set certificate verify locations:
* CAfile: /etc/ssl/certs/ca-certificates.crt
CApath: /etc/ssl/certs
* TLSv1.2 (OUT), TLS header, Certificate Status (22):
* TLSv1.2 (OUT), TLS handshake, Client hello (1):
* TLSv1.2 (IN), TLS handshake, Server hello (2):
* TLSv1.2 (IN), TLS handshake, Certificate (11):
* TLSv1.2 (IN), TLS handshake, Server key exchange (12):
* TLSv1.2 (IN), TLS handshake, Server finished (14):
* TLSv1.2 (OUT), TLS handshake, Client key exchange (16):
* TLSv1.2 (OUT), TLS change cipher, Client hello (1):
* TLSv1.2 (OUT), TLS handshake, Finished (20):
* TLSv1.2 (IN), TLS change cipher, Client hello (1):
* TLSv1.2 (IN), TLS handshake, Finished (20):
* SSL connection using TLSv1.2 / ECDHE-RSA-AES128-GCM-SHA256
* ALPN, server accepted to use http/1.1
* Server certificate:
* subject: CN=foo.bar.com
* start date: Apr 13 00:50:56 2016 GMT
* expire date: Apr 13 00:50:56 2017 GMT
* issuer: CN=foo.bar.com
* SSL certificate verify result: self signed certificate (18), continuing anyway.
> GET / HTTP/1.1
> Host: 10.2.78.7
> User-Agent: curl/7.47.1
> Accept: */*
>
< HTTP/1.1 404 Not Found
< Server: nginx/1.11.1
< Date: Mon, 18 Jul 2016 21:02:59 GMT
< Content-Type: text/html
< Transfer-Encoding: chunked
< Connection: keep-alive
< Strict-Transport-Security: max-age=15724800; includeSubDomains; preload
<
<span>The page you're looking for could not be found.</span>
* Connection #0 to host 10.2.78.7 left intact
Server-side HTTPS enforcement
By default the controller redirects (301) to HTTPS if TLS is enabled for that ingress . If you want to disable that behaviour globally, you can use ssl-redirect: "false" in the NGINX config map.
To configure this feature for specfic ingress resources, you can use the ingress.kubernetes.io/ssl-redirect: "false" annotation in theparticular resource.
HTTP Strict Transport Security
HTTP Strict Transport Security (HSTS) is an opt-in security enhancement specified through the use of a special response header. Once a supported browser receives this header that browser will prevent any communications from being sent over HTTP to the specified domain and will instead send all communications over HTTPS.
By default the controller redirects (301) to HTTPS if there is a TLS Ingress rule.
To disable this behavior use hsts=false in the NGINX config map.
Automated Certificate Management with Kube-Lego
Kube-Lego automatically requests missing certificates or expired from Let's Encrypt by monitoring ingress resources and its referenced secrets. To enable this for an ingress resource you have to add an annotation:
kubectl annotate ing ingress-demo kubernetes.io/tls-acme="true"
To setup Kube-Lego you can take a look at this full example. The first version to fully support Kube-Lego is nginx Ingress controller 0.8.
Exposing TCP services
Ingress does not support TCP services (yet). For this reason this Ingress controller uses the flag --tcp-services-configmap to point to an existing config map where the key is the external port to use and the value is <namespace/service name>:<service port>
It is possible to use a number or the name of the port.
The next example shows how to expose the service example-go running in the namespace default in the port 8080 using the port 9000
apiVersion: v1
kind: ConfigMap
metadata:
name: tcp-configmap-example
data:
9000: "default/example-go:8080"
Please check the tcp services example
Exposing UDP services
Since 1.9.13 NGINX provides UDP Load Balancing.
Ingress does not support UDP services (yet). For this reason this Ingress controller uses the flag --udp-services-configmap to point to an existing config map where the key is the external port to use and the value is <namespace/service name>:<service port>
It is possible to use a number or the name of the port.
The next example shows how to expose the service kube-dns running in the namespace kube-system in the port 53 using the port 53
apiVersion: v1
kind: ConfigMap
metadata:
name: udp-configmap-example
data:
53: "kube-system/kube-dns:53"
Please check the udp services example
Proxy Protocol
If you are using a L4 proxy to forward the traffic to the NGINX pods and terminate HTTP/HTTPS there, you will lose the remote endpoint's IP addresses. To prevent this you could use the Proxy Protocol for forwarding traffic, this will send the connection details before forwarding the acutal TCP connection itself.
Amongst others ELBs in AWS and HAProxy support Proxy Protocol.
Please check the proxy-protocol example
Custom errors
In case of an error in a request the body of the response is obtained from the default backend. Each request to the default backend includes two headers:
X-Codeindicates the HTTP codeX-Formatthe value of theAcceptheader
Using this two headers is possible to use a custom backend service like this one that inspect each request and returns a custom error page with the format expected by the client. Please check the example custom-errors
NGINX status page
The ngx_http_stub_status_module module provides access to basic status information. This is the default module active in the url /nginx_status.
This controller provides an alternitive to this module using nginx-module-vts third party module.
To use this module just provide a config map with the key enable-vts-status=true. The URL is exposed in the port 8080.
Please check the example example/rc-default.yaml
To extract the information in JSON format the module provides a custom URL: /nginx_status/format/json
Disabling NGINX ingress controller
Setting the annotation kubernetes.io/ingress.class to any value other than "nginx" or the empty string, will force the NGINX Ingress controller to ignore your Ingress. Do this if you wish to use one of the other Ingress controllers at the same time as the NGINX controller.
Debug & Troubleshooting
Using the flag --v=XX it is possible to increase the level of logging.
In particular:
--v=2shows details usingdiffabout the changes in the configuration in nginx
I0316 12:24:37.581267 1 utils.go:148] NGINX configuration diff a//etc/nginx/nginx.conf b//etc/nginx/nginx.conf
I0316 12:24:37.581356 1 utils.go:149] --- /tmp/922554809 2016-03-16 12:24:37.000000000 +0000
+++ /tmp/079811012 2016-03-16 12:24:37.000000000 +0000
@@ -235,7 +235,6 @@
upstream default-echoheadersx {
least_conn;
- server 10.2.112.124:5000;
server 10.2.208.50:5000;
}
I0316 12:24:37.610073 1 command.go:69] change in configuration detected. Reloading...
--v=3shows details about the service, Ingress rule, endpoint changes and it dumps the nginx configuration in JSON format--v=5configures NGINX in debug mode
These issues were encountered in past versions of Kubernetes:
- make setup-files.sh file in hypercube does not provide 10.0.0.1 IP to make-ca-certs, resulting in CA certs that are issued to the external cluster IP address rather then 10.0.0.1 -> this results in nginx-third-party-lb appearing to get stuck at "Utils.go:177 - Waiting for default/default-http-backend" in the docker logs. Kubernetes will eventually kill the container before nginx-third-party-lb times out with a message indicating that the CA certificate issuer is invalid (wrong ip), to verify this add zeros to the end of initialDelaySeconds and timeoutSeconds and reload the RC, and docker will log this error before kubernetes kills the container.
- To fix the above, setup-files.sh must be patched before the cluster is inited (refer to https://github.com/kubernetes/kubernetes/pull/21504)
Limitations
- Ingress rules for TLS require the definition of the field
host
NGINX notes
Since gcr.io/google_containers/nginx-slim:0.8 NGINX contains the next patches:
-
Dynamic TLS record size nginx__dynamic_tls_records.patch NGINX provides the parameter
ssl_buffer_sizeto adjust the size of the buffer. Default value in NGINX is 16KB. The ingress controller changes the default to 4KB. This improves the TLS Time To First Byte (TTTFB) but the size is fixed. This patches adapts the size of the buffer to the content is being served helping to improve the perceived latency. -
Add SPDY support back to Nginx with HTTP/2 nginx_1_9_15_http2_spdy.patch At the same NGINX introduced HTTP/2 support for SPDY was removed. This patch add support for SPDY wichout compromising HTTP/2 support using the Application-Layer Protocol Negotiation (ALPN) or Next Protocol Negotiation (NPN) Transport Layer Security (TLS) extension to negotiate what protocol the server and client support
openssl s_client -servername www.my-site.com -connect www.my-site.com:443 -nextprotoneg ''
CONNECTED(00000003)
Protocols advertised by server: h2, spdy/3.1, http/1.1
Documentation
¶
There is no documentation for this package.
Source Files
Directories