Warning: Draft Content

This wiki is closed and migrated to Cloud Native Deployment - for the last version check the history


Cloud Native Deployment#AmazonAWS


  • No labels

19 Comments

  1. I am unable to find the AMIs (ami-cd0f5cb6 and ami-c0c964ba) you have mentioned here, in AWS.

  2. Sorry, I made them private - as I realized my private key were in them - will update - anyway until Rancher 2 is vetted - you will need to do all the manual installs (could be 2 sh scripts between) - install docker, install rancher server, get rancher client auth-key (manual), register rancher client docker container, install helm, get kubectl key (manual), install kubectl - here is where the AMI stops.

    Ideally we get this into a fully automated sh script when Rancher 2 is out - and add this to the spot bootstrap as a cloud-init script.

    Do the following until you get to the undercloud - finished section - should take about 5-10 min from a clean Ubuntu 16.04

    using ami-cd0f5cb6

    ONAP on Kubernetes#QuickstartInstallation

    from spot on the EC2 service

    https://console.aws.amazon.com/ec2sp/v1/spot/launch-wizard?region=us-east-1 

    Exported json template - to see the id


    1. I was able to install Rancher 1.6.10 on AWS r4x2 instance, and then follow the steps for OOM installation as you mentioned.  The deployment was successful with the core components but ran into issues with Clamp, AAF, VFC, and DCAEGEN2 components.  I am aware that issues have been raised elsewhere in the wiki on these components, and that they are being tracked in JIRA. 

      1. I made another attempt to deploy with 'master' and 'amsterdam' branches using the cd.sh utility.

        For Amsterdam branch all components (minus DCAE) deployed, but only AAF component still had issues running.  When logging into the portal through vnc-portal, I could not get the SDC component to function properly.  It gives me a screen with user and category management options as shown below:

        For 'Master' branch I had trouble with pulling the docker image for vfc-ztevnfmdriver as shown below:

        docker pull nexus3.onap.org:10001/onap/vfc/ztevnfmdriver:v1.0.2
        Pulling repository nexus3.onap.org:10001/onap/vfc/ztevnfmdriver
        Error: image onap/vfc/ztevnfmdriver:v1.0.2 not found
        1. Rajeev,

             Very nice, yes there are known issues with aaf, vfc and kube2msb and periodically with sdc and sdnc.

             I will paste the jira.

             If the rest of your pods are up then you are good to go and are aligned with all the other types of deployment.

             That SDC screen means you are using the wrong user - see the page Running the ONAP Demos

             /michael

  3. Hi Michael O'Brien


     I am beginner to ONAP.
     So please forgive my ignorance.
     Thanks for the tutorial.I am able to install Onap in AWS environment. 

     My current scenario is I do not have any Openstack environment.

     I wanted to run some use cases on my current ONAP environment in AWS.

     As I went through the use cases almost all of them needs Openstack credentials to run the heat templates.

     Could you please guide me what can I do under my current ONAP environment in AWS.
     I could not find tutorials for what to do next after installing ONAP using Kubernetes in AWS.

     There are 90 pods that are installed among them aai and vfc has issues.

    I would be very great full if I can get some guide to run some use cases in my current environment in AWS.

     

    Thank you in advance.


    Best regards,

    Santosh 

    1. You will want to join the Multicloud project - we are developing plugins to support VNFs on Azure and AWS (no openstack required)

      Multi VIM/Cloud Project

      We meet Mondays at 2000 EST (GMT-5)


      1. Hi Michael O'Brien,


        Thank you very much for the suggestion.

        I would love to contribute from my side.

        I need another favor from you.

        I have a ONAP installed AWS using Kubernetes.

        However I do not have any openstack environment to run the use cases that needs openstack.

        Could you suggest me some use cases that i can test under my current ONAP system in Kubernetes(AWS).

        Best Regards,

        Santosh

  4. Hi Michael O'Brien

    Sorry for questions after questions.

    Is it possible to install openstack(Devstack) on other AWS instance and run the use cases.

    I want everything to be on AWS.

    Are  there any issues on installing devstack inside AWS instance.

    Best Regards,

    Santosh

    1. Yes, just have not done that myself yet - openstack on kubernetes is one option


      1. Sorry, another newbie question.


        How about a setup where ONAP (kubernetes version) is installed on AWS EC2, but deploying VNFs on Rackspace OpenStack?  Is such a setup possible without too many changes to the demo instructions?


        Perhaps my real question is which config file does ONAP read to figure out where to instantiate the VNFs such as vFW?  The vfw.env looks like it only describes how to create the vFW VM, but doesn't describe the OpenStack account in which to create the vFW VW.


        1. Your tenant is defined in onap_parameters.yaml

          master is simple https://git.onap.org/oom/tree/kubernetes/config/onap-parameters.yaml

          amsterdam is extensive (includes DCAEGEN2 heat parameters)

          https://git.onap.org/oom/tree/kubernetes/config/onap-parameters.yaml?h=amsterdam

  5. Hi Michael O'Brien,

    I was trying to run the onap on the AWS, but when I git clone the oom, i didn't find the oneclick directory in it.  Which has the createAll.bash to launch all the pods.

    Though the cd.sh does use the same and i see that none of the onap pods are launched post the completion of the cd.sh. All the docker files are downloaded and that's it.

    Is there a break in the master branch or it's supposed to be done differently.


    Thanks & Regards,

    Suraj

    1. master is retrofitted for helm install - createAll.bash is gone

      retry the rancher install and cd scripts and you should be fine

  6. Hi Michael O'Brien,

    I setup ONAP amsterdam on AWS. all modules of ONAP installed and run successfully. when i am trying to do ete health check, health check is failing for DCAE but all services of DCAE is running successfully. could you please help me to understand is something break. 

    Also when i am trying to install demo through ./demo_k8s.sh init. this command is sowing all failed run not able to figure out what wrong. please help me. 

    Thanks,

    Vijendra


    1. Healtcheck passes now in beijing - I have not been on amsterdam for a while

      http://jenkins.onap.info/job/oom-cd-master/2875/console

  7. Hi Micheal,

    I have successfully installed Amsterdam on AWS all pods in kubernetes running as expected. Now trying to play with vFWCL usecase for the same i installed openstack on different AWS instance but when i see the heat template provided in vFWCL usecase seem so many network address (e.g. onap_private, private_unprotected, public etc.) are used. i changed those addressed as per my openstack instance but seem not working as expected. could you please help me to figure out any changes would be require in heat template.

    Thanks!

    1. These are detailed in RTD and subject to change as we consolidate all the entries in values.yaml

      http://onap.readthedocs.io/en/latest/submodules/oom.git/docs/oom_quickstart_guide.html

      There is a JIRA to add comments per line like we did in Dec when vetting the vFW use case - I'll find the JIRA

      https://git.onap.org/oom/tree/kubernetes/onap/values.yaml