This is an example of how to run an ONAP sdnc-image with a minimal setup with SDN-R configuration with docker-compose. With this deployment you are able run the ODLUX gui within the SDNC which is then reachable on http://localhost:8181/odlux/index.html.
Prerequisites
- vm (4 vCPU/ 4 gig RAM/ 40Gig HDD)
- ubuntu (e.g. Ubuntu server 20.04.2 LTS)
- docker
- docker-compose (v2 recommended)
Deployment
So we have only 2 files in our directory. The docker-compose.yml itself and the .env file for storing all parameters we may like to change.
docker-compose.yml (minimal):
services: sdnc-web: image: ${SDNC_WEB_IMAGE} container_name: sdnc-web ports: - 8080:8080 environment: - SDNRHOST=sdnr depends_on: sdnr: condition: service_healthy healthcheck: test: "curl --silent --fail http://localhost:8080/ready > /dev/null || exit 1" interval: 10s timeout: 10s retries: 45 sdnr: image: ${SDNC_IMAGE} container_name: sdnr ports: - 8181:8181 environment: - A1_ADAPTER_NORTHBOUND=false - CCSDK_REPLICAS=0 - DOMAIN="" - ENABLE_OAUTH=false - ENABLE_ODL_CLUSTER=false - ODL_ADMIN_PASSWORD=${ADMIN_PASSWORD} - SDNC_REPLICAS=0 - SDNRDBURL=http://sdnrdb:9200 - SDNRDM=false - SDNRINIT=true - SDNRONLY=true - SDNRWT=true - SDNR_NETCONF_CALLHOME_ENABLED=true - JAVA_OPTS=-Xms256m -Xmx4g healthcheck: test: [ "CMD", "wget", "-nv", "-t1", "--spider", "http://localhost:8181/ready" ] interval: 10s timeout: 10s retries: 45 depends_on: - sdnrdb sdnrdb: image: ${SDNRDB_IMAGE} container_name: sdnrdb environment: - discovery.type=single-node
.env file:
SDNC_IMAGE=nexus3.onap.org:10001/onap/sdnc-image:2.6.1 SDNC_WEB_IMAGE=nexus3.onap.org:10001/onap/sdnc-web-image:2.6.1 ADMIN_PASSWORD=admin SDNRDB_IMAGE=docker.elastic.co/elasticsearch/elasticsearch-oss:7.9.3
The sdnc image version (here latest) you can choose from this list here: https://nexus3.onap.org:10001/v2/onap/sdnc-image/tags/list.
version | ONAP release name | odl-base |
---|---|---|
2.8.X | Oslo | potassium-sr2 |
2.7.X | - | |
2.6.X | Montreal | argon-sr2 |
2.5.X | London | chlorine |
2.4.X | Kohn | sulfur |
2.3.X | Jakarta | phosphorus |
2.2.X | Istanbul | silicon-SR1 |
2.1.X | Honolulu | aluminium-SR1/3 |
2.0.X | Guilin | sodium-SR3/4 |
1.8.X | Frankfurt | |
1.7.X | El Alto |
So after you started this deployment with
$ docker-compose up -d /docker compose up -d (depends on your docker compose version)
you should be able to access the ODLUX gui via you webbrowser on http://localhost:8181/odlux/index.html. Login credentials are username "admin" and password the one you have in your .env file, so by default also "admin".
Troubleshooting
Problem: docker container is crashing
check the status of the container
$ docker ps -a | grep sdnr
If the status is "up ..." then you may have to wait a little bit more.
If the status is exited, check the stdout logs of the container
$ docker logs sdnr
Problem: ODLUX is not coming up/ 401 Unauthorized response:
Step 1: check ready service
$ curl -v http://localhost:8181/ready $ curl -v http://localhost:8080/ready
both should return response code 200
< HTTP/1.1 200 OK < Server: nginx < Date: Mon, 20 Sep 2021 07:00:29 GMT < Transfer-Encoding: chunked < Connection: keep-alive < X-Frame-Options: SAMEORIGIN < * Connection #0 to host localhost left intact
If this does not return a 200 something is wrong on the sdnc or sdnc-web side so you have to follow step 2. If it returns 200 and odlux is still not reachable also go to step 2.
Step2: check odl console
$ docker exec -ti sdnr bash bash-5.0$ /opt/opendaylight/bin/client
if this is not accessable go to step 3:
if this is accessable do:
$ opendaylight-user@root> bundle:list | grep -iv active
the output should be something like this:
START LEVEL 100 , List Threshold: 50 ID │ State │ Lvl │ Version │ Name ────┼──────────┼─────┼────────────────────────────────────────┼─────────────────────────────────────────────────────── 19 │ Resolved │ 80 │ 8.1.2 │ ODL :: odlparent :: bcprov-framework-ext, Hosts: 0 20 │ Resolved │ 80 │ 8.1.2 │ ODL :: odlparent :: bcpkix-framework-ext, Hosts: 0 111 │ Resolved │ 80 │ 4.3.1 │ Apache Karaf :: Diagnostic :: Boot
if there is one package on Failure state pls go to step 3. If not go on with
$ opendaylight-user@root> bundle:list | grep -i ccsdk
If this list has some services still on GracePeriod, pls wait some more time and try it again (Can take up to two minutes). If this list is empty it looks like no sdnr feature was started. Then you have to check the features config file.
$ opendaylight-user@root> less etc/org.apache.karaf.features.cfg | grep featuresBoot
there you should see something like this:
featuresBoot = 3a3f1168-5df1-4fcc-a03f-6253067eaa1a,odl-restconf-all,odl-netconf-topology,odl-mdsal-all,odl-mdsal-apidocs,odl-daexim-all,odl-restconf-nb-rfc8040,sdnr-wt-feature-aggregator-devicemanager-base,sdnr-wt-feature-aggregator-devicemanager featuresBootAsynchronous=false
If there are no sdnr features inside of this list, please check your environment variables inside of the container. If these are also okay, then you have to check the startODL.sh script for the install_sdnr_wt() function call, because this is then not called. If only odlux features are missing you may have set the SDNRDM env var to true. This is possible, but needs an extended setup which is not part of this guide.
Step 3: Search the log for failure inside of the container
$ docker exec -ti sdnr bash bash-5.0$ less /opt/opendaylight/data/log/karaf.log
Here you may find a resolve error or something similar. If this happens with a non release image (X.X.X-STAGING-xxx) pls try to use a release (X.X.X) first.