User Tools

Site Tools


OPNFV Brahmaputra User Guide:

Installation Guide

Post installation guide

User Guide:

OPNFV requirements:

  • Installers shall support the deployment of OpenStack with High-Availability (for those components that support it in Liberty) on 3 or more control nodes. Functest tests should be able to verify that the HA is enabled and functional.
  • Hitless hardware upgrade: Increase size of a deployment in terms of compute nodes (add additional compute nodes) without service interruption. This requirement doesn't mandate upgrade/increasing the size of the control node cluster.
  • Automatically populate discovered servers into install tool (reduce user-intervention to a minimum).
  • Installers should support a common configuration file per platform/role, so that the installed OS can be customized for hardware and role.
  • Neutron DHCP servers should be configured in HA per tenant.
  • SDN Controller layer 3 forwarding support.
  • Support layer 1/2 networking configuration.
  • Support logical networks for target system.
  • Support L3-neutron agent as an option for L3.
  • Provide isolated compute node resources for CEPH OSD.
  • Installers should track/control all versions of all components pulled from external sources (user should be able to identify the versions and origins of all software components deployed).
  • Common ability to input site, topology, and server information.
  • User-configurable parameters available via config files.
  • Allow assignment of different roles to servers, so that hardware and software can be configured according to the role.
  • Deployment tool to provide for automatic device discovery.
  • Installers should configure NTP servers on the servers for clock synchronization.
  • Installers to be agnostic to type of hard drives used.
  • Installers should install ssh keys on servers so that key-based login can be used for administration.
  • Ability to install with upstream artifacts.
  • Installers should supply a script or set of scripts ("")
  • Installers should support offline deployment. Jump host may have Internet access, but the installers should support offline installation on target hosts during the deployment phase (either manually or automatically).

Access the openstack dashboard:

Users can access the dashboard from the jumpshot directly. If you are working on jumpshot then please open a web browser on the jumpshot and start accessing the openstack dashboard.

http:<vip>/ for Intel pod5 credentials: admin / openstack === Command line: === * source ./cloud/admin-openrc * openstack role list * openstack user list * glance image-list * nova list * keystone tenanant-list ==== Troubleshooting: ==== JOID is combination of MAAS and JUJU tools which do the OPNFV deployment on bare metal as well as on virtual machine. 1. MAAS deployment In case you are troubleshoot the deployment of then please look into logs during run time of any errors. Possible errors could be. a) Hardware does not support the virtualization. Please enable the virtualization in BIOS and rerun b) ERROR error: Failed to delete vol bootstrap.img error: cannot unlink file '/var/lib/libvirt/images/bootstrap.img': Success error: Failed to delete vol bootstrap.img error: cannot unlink file '/var/lib/libvirt/images/bootstrap.img': Success subprocess.CalledProcessError: Command 'virsh -c qemu:/system vol-delete –pool default

      bootstrap.img' returned non-zero exit status 1
   solution: There could be some corruption on the images at /var/lib/libvirt/images/ please remove all 
      image file from the directory /var/lib/libvirt/images/ and reboot the server then start 
   c) timeout issue. 
      Make sure you have good internet connection and your server is not behind proxy and try it again. 
2. JUJU deployment
   You may encounter issues in your lab environment. Below could be issues you may encounter
   a) Right variable are not passed as part of deployment procedure.
      ./ -o liberty -s odl -t ha -l intelpod5 -f none
   b) if you have setup maas not with then you could encounter hang on ./
      juju status command may hang because correct MAAS API keys are not listed in envrinments.yaml 
      or environments.yaml does not exist in current working directory.
      Solution: Please make sure you have environments.yaml file under joid/ci directory and correct   
      MAAS API key has been listed.
   c) deployment stuck at long time and timeout:
      use command juju status --format=tabular and make sure all services container got an ip and they are executing some code. There is no service in the error state.
3. Openstack deployment
joid/b_userguide.txt · Last modified: 2016/02/11 05:07 by Narinder Gupta