docker-netbootxyz/kubernetes

Kubernetes Deployment example

This is an example of one possibilty how to deploy netboot.xyz to a OKD/OpenShift cluster.
You can also use it for a Kubernetes cluster, but in this case you will add your ingress configuration instead of the route.

Edit PVC config

First edit both pvc (pvc.yaml and pvc-config.yaml) configs, so that they reference your storage-class.

Edit Route config (for OKD/OpenShift)

Edit the host to your needs.

spec:
  host: pxeboot.apps.cluster.local

Optional edit PXE-Bootserver-Conf

Optionaly you can edit the pxe-bootserver-conf.yaml for your needs.

Deploy to Kubernetes or OKD/OpenShift

create a namespace

kubectl create ns network

deploy

kubectl -n network apply -f pvc.yaml
kubectl -n network apply -f pvc-config.yaml
kubectl -n network apply -f pxe-bootserver-conf.yaml
kubectl -n network apply -f serviceaccount.yaml
kubectl -n network apply -f deployment.yaml
kubectl -n network apply -f route.yaml
kubectl -n network apply -f service.yaml

CNI settings for the TFTP-Server

Please notice the service configuration for the service:

• svc-pxboot this is from type: NodePort.
  This is important, because after a TFTP-client requests a file,
  the TFTP-Server will initiate a new connection and send data back to the client over this new connection.
  So you must configure your CNI do not use source nat (SNAT) for this connections !
  CNI configs:
• calico SNAT-Config
• cilium SNAT-Config

Check if netboot.xyz is running

Check Deployment, Service and Pod

kubectl -n network get all

NAME                                        READY   STATUS    RESTARTS   AGE
pod/pxe-bootserver-ds-5559fd7-4ncjb         1/1     Running   0          2d23h

NAME                               TYPE        CLUSTER-IP       EXTERNAL-IP   PORT(S)           AGE
service/svc-pxe-bootserver         ClusterIP   10.20.255.66     <none>        80/TCP,3000/TCP   15d
service/svc-pxeboot                NodePort    10.20.255.51     <none>        69:30936/UDP      15d

NAME                                   READY   UP-TO-DATE   AVAILABLE   AGE
deployment.apps/pxe-bootserver-ds      1/1     1            1           3d7h

NAME                                              DESIRED   CURRENT   READY   AGE
replicaset.apps/pxe-bootserver-ds-5559fd7         1         1         1       2d23h

NAME                                                  HOST/PORT                           PATH   SERVICES                   PORT    TERMINATION   WILDCARD
route.route.openshift.io/pxe-bootserver-route         pxeboot.apps.cluster.local                 svc-pxe-bootserver         3000    edge          None

Check the logs of the Pod

kubectl -n network logs pxe-bootserver-ds-5559fd7-4ncjb

            _   _                 _
 _ __   ___| |_| |__   ___   ___ | |_  __  ___   _ ____
| '_ \ / _ \ __| '_ \ / _ \ / _ \| __| \ \/ / | | |_  /
| | | |  __/ |_| |_) | (_) | (_) | |_ _ >  <| |_| |/ /
|_| |_|\___|\__|_.__/ \___/ \___/ \__(_)_/\_\__,  /___|
                                             |___/
2024-12-03 16:40:33,309 INFO Set uid to user 0 succeeded
2024-12-03 16:40:33,322 CRIT could not write pidfile /supervisord.pid
2024-12-03 16:40:34,326 INFO spawned: 'syslog-ng' with pid 13
2024-12-03 16:40:34,330 INFO spawned: 'nginx' with pid 14
2024-12-03 16:40:34,333 INFO spawned: 'webapp' with pid 15
2024-12-03 16:40:34,336 INFO spawned: 'dnsmasq' with pid 16
2024-12-03 16:40:34,340 INFO spawned: 'messages-log' with pid 17
2024-12-03 16:40:34,348 WARN exited: messages-log (exit status 1; not expected)
2024-12-03 16:40:34,367 WARN exited: syslog-ng (exit status 2; not expected)
2024-12-03 16:40:35,714 INFO spawned: 'syslog-ng' with pid 28
2024-12-03 16:40:35,715 INFO success: nginx entered RUNNING state, process has stayed up for > than 1 seconds (startsecs)
2024-12-03 16:40:35,715 INFO success: webapp entered RUNNING state, process has stayed up for > than 1 seconds (startsecs)
2024-12-03 16:40:35,715 INFO success: dnsmasq entered RUNNING state, process has stayed up for > than 1 seconds (startsecs)
2024-12-03 16:40:35,718 INFO spawned: 'messages-log' with pid 29
2024-12-03 16:40:35,726 WARN exited: messages-log (exit status 1; not expected)
2024-12-03 16:40:35,754 WARN exited: syslog-ng (exit status 2; not expected)
2024-12-03 16:40:37,760 INFO spawned: 'syslog-ng' with pid 30
2024-12-03 16:40:37,763 INFO spawned: 'messages-log' with pid 31
2024-12-03 16:40:37,771 WARN exited: messages-log (exit status 1; not expected)
2024-12-03 16:40:37,802 WARN exited: syslog-ng (exit status 2; not expected)
2024-12-03 16:40:40,938 INFO spawned: 'syslog-ng' with pid 36
2024-12-03 16:40:40,941 INFO spawned: 'messages-log' with pid 37
2024-12-03 16:40:40,950 WARN exited: messages-log (exit status 1; not expected)
2024-12-03 16:40:40,950 INFO gave up: messages-log entered FATAL state, too many start retries too quickly
2024-12-03 16:40:40,981 WARN exited: syslog-ng (exit status 2; not expected)
2024-12-03 16:40:41,983 INFO gave up: syslog-ng entered FATAL state, too many start retries too quickly

In the current version we have 2 bugs, which have no impact for the netboot.xyz server.

syslog-ng       must be reconfigured or have to be removed
messages-log    must be reconfigured or have to be removed

DHCP-Options for TFTP

This deployment will need the following DHCP-Options:

• Option 66 Boot Server Hostname: tftp.svc-pxeboot.network.cluster.local
• Option 67 Bootfile Name: netboot.xyz.efi (for UEFI-Boot)
• Option 67 Bootfile Name: netboot.xyz-undionly.kpxe (for BIOS-Boot)