...
The instructions to create an ONAP installation using the OOM Rancher/Kubernetes approach are in the ONAP wiki site (be sure to select the Casablanca version of the instructions). Once installed, there are further instructions on deploying ONAP at this wiki page. To install the development image rather than the nexus3 image, we must update parameter values in the Helm chart for SDNC in the OOM repository, shown here.
Code Block |
---|
% ls -F git/oom/kubernetes/sdnc
charts/ Chart.yaml Makefile requirements.lock requirements.yaml resources/ sdnc-prom/ templates/ values.yaml
|
Override file for the SDNC values.yaml file
The simplest way to override the values is to copy the entire values.yaml file and modify the relevant parameters. The new values are shown here in a separate override-sdnc.yaml file. We identify the repository with the source image name and tag, create a cluster of three ODL members, and create a redundant MySQL deployment of two instances.
...
#################################################################
# Application configuration defaults.
#################################################################
# application images
repository: nexus3.onap.org:10001
repositoryOverride: registry.hub.docker.com
pullPolicy: Always
#image: onap/sdnc-image:1.4.1
image: ft3e0tab7p92qsoceonq/oof-pci-sdnr:1.4.2-SNAPSHOT
...
mysql:
nameOverride: sdnc-db
service:
name: sdnc-dbhost
internalPort: 3306
nfsprovisionerPrefix: sdnc
sdnctlPrefix: sdnc
persistence:
mountSubPath: sdnc/mysql
enabled: true
disableNfsProvisioner: true
replicaCount: 2
geoEnabled: false
...
# default number of instances
replicaCount: 3
...
Override file for the ONAP values.yaml file
By default, the OOM Rancher/Kubernetes script installs all of the components, which we do not need for the proof-of-concept. We identify which components to install by copying the ~/git/oom/kubernetes/onap/values.yaml file into a separate "override" file and changing "enabled: true" to "enabled: false" for the unneeded components. Currently, these are the selected components.
aaf | false |
aai | true |
appc | false |
clamp | false |
cli | false |
consul | false |
contrib | false |
dcaegen2 | false |
dmaap | true |
esr | false |
log | true |
sniro-emulator | true |
oof | true |
msb | false |
multicloud | false |
nbi | false |
policy | true |
pomba | false |
portal | true |
robot | true |
sdc | true |
sdnc | true |
so | true |
uui | false |
vfc | false |
vid | false |
vnfsdk | false |
Command to install ONAP with the development image
Following the guidelines at the OOM wiki page, I use this command to install ONAP with the desired configuration.
Code Block |
---|
cd ~/git/oom/kubernetes
sudo helm deploy demo ./onap --namespace onap -f ~/oof-pci/override-onap.yaml -f ~/oof-pci/override-sdnc.yaml |
The parameter "demo" is used to preface each ONAP component with "demo-" so we have "demo-sdnc," for example. The "./onap" parameter instructs helm to use that directory to guide the deployment. The "–namespace onap" parameter causes ONAP to be deployed into the kubernetes namespace "onap." The "-f ~/oof-pci/override-onap.yaml -f ~/oof-pci/override-sdnc.yaml" parameters instruct helm to override the parameters in the ~/git/oom/kubernetes/onap/values.yaml file with the values in the files following the "-f" option. There can be a series of override files, and the last file takes precedence.
Commands to install the development image
If there is already an instance SDNC installed, it must be deleted before installing a new version. I use these commands.
Code Block |
---|
helm del demo-sdnc --purge
kubectl get persistentvolumes -n onap | grep demo-sdnc | sed 's/\(^[^ ]\+\).*/kubectl delete persistentvolumes -n onap \1/'
kubectl get persistentvolumeclaims -n onap | grep demo-sdnc | sed 's/\(^[^ ]\+\).*/kubectl delete persistentvolumeclaims -n onap \1/'
kubectl get secrets -n onap | grep demo-sdnc | sed 's/\(^[^ ]\+\).*/kubectl delete secrets -n onap \1/'
kubectl get clusterrolebindings -n onap | grep demo-sdnc | sed 's/\(^[^ ]\+\).*/kubectl delete clusterrolebindings -n onap \1/' |
The first command deletes SDNC but, despite the "–purge" option, some residual resources remain. The subsequent commands discovers those resources and generates commands that can be copied and pasted into your terminal session to be executed. If you know how to pipe a string into bash so it can be executed directly, kindly update this code. The "helm del..." command takes some time, so please be patient. Once SDNC has been deleted, you can install the new version using the commands in the previous section.
Accessing SDNC/SDNR
Now that SDNC/SDNR is deployed, how can you access it? I use this sequence of commands. First:
Code Block |
---|
% kubectl get pods -n onap -o wide | grep NODE && kubectl get pods -n onap -o wide | grep sdnc
NAME READY STATUS RESTARTS AGE IP NODE
demo-sdnc-controller-blueprints-694b7ff9d-gmrtd 1/1 Running 0 20h 10.42.42.8 sb4-k8s-1
demo-sdnc-controller-blueprints-db-0 1/1 Running 0 20h 10.42.100.157 sb4-k8s-1
demo-sdnc-nengdb-0 1/1 Running 0 20h 10.42.101.202 sb4-k8s-3
demo-sdnc-network-name-gen-7fc56878b6-sz8ps 1/1 Running 0 20h 10.42.55.243 sb4-k8s-3
demo-sdnc-sdnc-0 2/2 Running 0 20h 10.42.105.225 sb4-k8s-4
demo-sdnc-sdnc-1 2/2 Running 0 20h 10.42.11.48 sb4-k8s-1
demo-sdnc-sdnc-2 2/2 Running 0 20h 10.42.9.208 sb4-k8s-2
demo-sdnc-sdnc-ansible-server-7ddf4c54dd-vq877 1/1 Running 0 20h 10.42.137.38 sb4-k8s-1
demo-sdnc-sdnc-db-0 2/2 Running 0 20h 10.42.119.112 sb4-k8s-3
demo-sdnc-sdnc-db-1 2/2 Running 0 20h 10.42.26.168 sb4-k8s-4
demo-sdnc-sdnc-dgbuilder-647d9bddb8-b2gxp 1/1 Running 0 20h 10.42.93.148 sb4-k8s-4
demo-sdnc-sdnc-dmaap-listener-f9c9fd74c-w42rq 0/1 Init:0/1 0 20h 10.42.38.155 sb4-k8s-3
demo-sdnc-sdnc-portal-6fcd6b8445-bhf48 1/1 Running 0 20h 10.42.249.112 sb4-k8s-4
demo-sdnc-sdnc-ueb-listener-849d6498b5-mf8pw 0/1 Init:0/1 0 20h 10.42.0.101 sb4-k8s-3
demo-so-so-sdnc-adapter-5b7787596d-bm9xn 1/1 Running 0 2d 10.42.170.141 sb4-k8s-1
% ping sb4-k8s-4
PING sb4-k8s-4 (10.31.1.79) 56(84) bytes of data.
64 bytes from sb4-k8s-4 (10.31.1.79): icmp_seq=1 ttl=64 time=0.505 ms |
We see that there are three instances of SDNC running and two instances of SDNC-DB and that they are deployed in different nodes, as expected. All of the pods have private IP addresses that are not accessible from outside the ONAP deployment, but demo-sdnc-sdnc-0 is installed in NODE sb4-k8s-4, which has IP address 10.31.1.79. We now enter this command.
Code Block |
---|
% kubectl get svc -n onap | grep NAME && kubectl get svc -n onap | grep sdnc
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
sdnc NodePort 10.43.141.133 <none> 8282:30202/TCP,8202:30208/TCP,8280:30246/TCP,8443:30267/TCP 20h
sdnc-ansible-server ClusterIP 10.43.41.91 <none> 8000/TCP 20h
sdnc-cluster ClusterIP None <none> 2550/TCP 20h
sdnc-dbhost ClusterIP None <none> 3306/TCP 20h
sdnc-dbhost-read ClusterIP 10.43.100.184 <none> 3306/TCP 20h
sdnc-dgbuilder NodePort 10.43.16.12 <none> 3000:30203/TCP 20h
sdnc-dmaap-listener ClusterIP None <none> <none> 20h
sdnc-portal NodePort 10.43.40.149 <none> 8843:30201/TCP 20h
sdnc-sdnctldb01 ClusterIP None <none> 3306/TCP 20h
sdnc-sdnctldb02 ClusterIP None <none> 3306/TCP 20h
sdnc-ueb-listener ClusterIP None <none> <none> 20h
so-sdnc-adapter ClusterIP 10.43.141.124 <none> 8086/TCP 2d |
SDNC is presenting a service at a NodePort that is accessible from outside the ONAP installation. PORT 8282:30202 means that port 30202 is accessible externally and maps to internal port 8282 (I'm not sure why 8282 rather than 8181; a port mapping from 8282 to 8181 may be set in a Dockerfile). Therefore, SDNC is listening at sb4-k8s-4:30202, or 10.31.1.79:30202. By creating a ssh tunnel to sb4-k8s-4 (described here), one can open a browser to localhost:30202/apidoc/explorer/index.html and see this.
Conclusion
Please feel free to edit this page to make corrections or improvements. Your assistance will be greatly appreciated.