diff options
Diffstat (limited to 'docs')
-rw-r--r-- | docs/developer_info/ETSI_CSIT_NFVO_VNFM.rst | 227 |
1 files changed, 227 insertions, 0 deletions
diff --git a/docs/developer_info/ETSI_CSIT_NFVO_VNFM.rst b/docs/developer_info/ETSI_CSIT_NFVO_VNFM.rst new file mode 100644 index 0000000000..3bc8393d56 --- /dev/null +++ b/docs/developer_info/ETSI_CSIT_NFVO_VNFM.rst @@ -0,0 +1,227 @@ +.. This work is licensed under a Creative Commons Attribution 4.0 International License. +.. http://creativecommons.org/licenses/by/4.0 +.. Copyright 2022 Ericsson Software Technologies + +SO ETSI CSIT (NFVO and VNFM) +============================ +This guide will go through the CSIT of the ETSI NFVO and VNFM. + +Ensure you have a healthy ONAP Deployment running. The following components will be required/used as part of this guide: + +- SO +- SDC +- AAI +- DMAAP +- Modeling + +What is ETSI? +------------- +The European Telecommunications Standards Institute (ETSI) produces globally-applicable standards +for Information and Communications Technologies. ETSI set out standards covering the functionalities +of the interfaces specified on the reference points, which use the acronym NFV-SOL +(standing for “NFV Solutions”). As of ONAPs Dublin release the SO SVNFM adapter supports +SOL003 standards for Create, Instantiate, Terminate and Delete operations with Granting, Subscription +and Notification. As of ONAP Honolulu release, the SO ETSI NFVO sub-component supports +SOL005 (NBI) and SOL003 (SBI) standards for Create, Instantiate, Terminate and Delete NS and VNF +(through the SOL003 Adapter) operations. + +How to Run CSIT Tests +--------------------- +The follow steps are to install and run on an Ubuntu 18.04 (desktop) installation. +Later versions of Ubuntu and Java may work for the tests. + +First pull the CSIT repo from Gerrit, either with or without the hooks + +.. code-block:: + + git clone "https://gerrit.onap.org/r/integration/csit" + +or + +.. code-block:: + + git clone "https://gerrit.onap.org/r/integration/csit" && (cd "csit" && mkdir -p .git/hooks && curl -Lo `git rev-parse --git-dir`/hooks/commit-msg https://gerrit.onap.org/r/tools/hooks/commit-msg; chmod +x `git rev-parse --git-dir`/hooks/commit-msg) + +Once this is downloaded a few more installations are required. + +Install pip (if required) + +.. code-block:: + + sudo apt install python-pip + +Install Robot Framework through pip + +.. code-block:: + + pip install robotframework + +Run Script + +Once all of this is done, then the tests should be run by calling the run-csit.sh script and giving it the location of our test folder (csit/plans/so/integration-etsi-testing). +From the csit projects root folder run the following command: + +.. code-block:: + + ./run-csit.sh plans/so/integration-etsi-testing + +This should successfully run the ETSI CSIT suite + +How to run tests againt specific SO versions +-------------------------------------------- +It is possible to run the ETSI CSIT suite against local docker images although it is not the default. +Through this method specific versions of SO can be tested. +There are two changes required to make this work. +The env file, located at [containing folder]/csit/plans/so/integration-etsi-testing/config/env, +first needs to be changed. The DOCKER_ENVIROMENT needs to be changed from "remote" to "local". +Also the TAG value might need to be changed. This Tag relates to the version of images being used. + +Secondly all of the required docker images must be present on system. + +This should be enough to run the ETSI CSIT test suite locally. + +ETSI NFVO Automated CSIT Tests High Level Scenarios +--------------------------------------------------- +**Step 1:** + +Perform Configuration / Setup Steps prior to running tests + +**Step 2:** + +Onboard SOL004 Package and SOL007 directly into ETSI Catalog using ROBOT framework + +**Step 3:** + +ETSI Catalog Gets Package from SDC Simulator (New: ETSI Catalog and Modeling ETSI Catalog DB will need to be set up and spun up for CSIT. May be some impact on SDC Simulator here) + +**Step 4:** + +ETSI Catalog Stores Package in its Modeling ETSI Catalog DB + +**Step 5:** + +ROBOT framework used to trigger NS LCM requests, e.g., INSTANTIATE NS + +**Step 6:** + +ETSI NFVO NS LCM gets required data from ETSI Catalog, e.g., Get and Parse NSD + +**Step 7:** + +If e.g., a CREATE NS task, ETSI NFVO NS LCM checks to see if already exists in ETSI NFVO DB + +**Step 8:** + +Create Generic VNF and connect to Service Instance in A&AI Simulator (May be some impact on A&AI Simulator here) + +**Step 9:** + +Instantiate VNF through SOL003 Adapter + +**Step 10:** + +SOL003 Adapter processes requests through A&AI and (May be some impact on A&AI Simulator here) + +**Step 11:** + +SOL003 Adapter processes requests through ETSI-Catalog + +**Step 12:** + +SOL003 Adapter sends notification to SOL003 NBI, etc. + +What are the tests doing? +------------------------- +There are three tests currently being run "Distribute Service Template", "Invoke Service Instantiation", +"Invoke NS Instantiation", "Delete NS Instance", "Invoke VNF Instantiation", "Delete VNF Instance" and +"Delete Service Instance". + +Distribute Service Template + +As the name would suggest the aim for the "Distribute Service Template" test is to distribute a service +template within the SDC controller pod. Once a http session of the SDC controller is created a post request +can be made to it. This post requests sends binary data from "distributeServiceTemplate.json". +This json file contains resources and artifacts required to distribute a service. Once this post request +is sent, the response status code is checked to see if it is 200. If the code is not equal to 200 then +the test is thought to be a failure. + +Invoke Service Instantiation + +The aim of the "Invoke Service Instantiation" test is to invoke the service distributed to the sdc controller +in the previous test. A http session of the api handler pod is created. This session is sent a post request +containing "serviceInstantiationRequest.json". Once this request is made the response is checked if it +a valid code is returned. A for loop is used to continually make calls to check the orchestration request, +to check the status of service instantiation. Only once this orchestration returns either a fail or success, +will we break out of the for loop.Once outside the for loop a final statement is used to check if service +has been successfully instantiated. + +Invoke NS Instance + +The aim of "Invoke NS Instantiation" test is to now instantiate the NS that relates to service in the +previous test. This test requires the ID of the service instance created in the previous test. If this is +not provided then the test will fail from the get go. Once again a http session of the api handler pod is +created. Similarly a post request using the json data within "nsInstantiationRequest.json". +Once this request is made if it returns a success code then the test moves on to a for loop. Within this +for a loop an orchestration request is made each time, when this request signals that either the instantiation +request has failed or fully succeeded then the loop is escaped. The test will either be a pass or fail depending +on this final orchestration request. + +Delete NS Instance + +This test will delete the NS Instance created in the previous test. Both the ID of the NS instance created +in the previous test and the service instance created in the test before that. If either of these values is +not provided then the test will fail. This test once again makes use of a session of the api handler pod. +A post request is made using the data from "nsDeleteRequest.json". Once this request is made if it returns +a success code then the test moves on to a for loop. Within this for a loop an orchestration request is made +each time, when this request signals that either the instantiation request has failed or fully succeeded +then the loop is escaped. The test will either be a pass or fail depending on this final orchestration request. + +Invoke VNF Instance + +The aim of "Invoke VNF Instantiation" test is to now instantiate the VNF that relates to service in +the previous test. This test requires the ID of the service instance created in the previous test. +If this is not provided then the test will fail from the get go. Once again a http session of the +api handler pod is created. Similarly a post request using the json data within "vnfInstantiationRequest.json". +Once this request is made if it returns a success code then the test moves on to a for loop. Within this +for a loop an orchestration request is made each time, when this request signals that either the instantiation +request has failed or fully succeeded then the loop is escaped. The test will either be a pass or fail +depending on this final orchestration request. + +Delete VNF Instance + +This test will delete the VNF Instance created in the previous test. Both the ID of the vnf instance created +in the previous test and the service instance created in the test before that. If either of these values is +not provided then the test will fail. This test once again makes use of a session of the api handler pod. +A post request is made using the data from "vnfDeleteRequest.json". Once this request is made if it returns +a success code then the test moves on to a for loop. Within this for a loop an orchestration request is made +each time, when this request signals that either the instantiation request has failed or fully succeeded then +the loop is escaped. The test will either be a pass or fail depending on this final orchestration request. + +Delete Service Instance + +This test will delete the service instance created in earlier test. To delete the service the ID of previously + created Service Instance is required, if this is not supplied then the test will fail before starting. + A post request is then made to the API handler containing data from "serviceDeleteRquest.json". + Once this request is made if it returns a success code then the test moves on to a for loop. + Within this for a loop an orchestration request is made each time, when this request signals that either + the instantiation request has failed or fully succeeded then the loop is escaped. The test will either be + a pass or fail depending on this final orchestration request. + +Troubleshooting +--------------- +There are a number of simple issues relating from Python and its libraries + +A correct installation of the robot framework to run our tests requiring python and the following pip libraries. + +- robotframework +- robotframework-extendedselenium2library +- robotframework-httplibrary +- robotframework-onap +- robotframework-requests +- robotframework-selenium2library + +To make sure each of the previous libraries is installed run the following command + +.. code-block:: + + pip -list
\ No newline at end of file |