|author||Nikos Karandreas <firstname.lastname@example.org>||2018-12-12 08:42:07 +0000|
|committer||Gerrit Code Review <email@example.com>||2018-12-12 08:42:07 +0000|
diff --git a/docs/release/installation/index.rst b/docs/release/installation/index.rst
index ded6a78..089bc55 100644
@@ -31,7 +31,9 @@ spec>.
When ODL is used as an SDN Controller in an OPNFV virtual deployment, ODL is
running on the OpenStack Controller VMs. It is therefore recommended to
-increase the amount of resources for these VMs.
+increase the amount of resources for these VMs. ODL is running in a separate
+VM in case of Fuel, thus, the below recommendation is not applicable when
+deploying the scenario on Fuel installer.
Our recommendation is to have 2 additional virtual cores and 8GB
additional virtual memory on top of the normally recommended
@@ -50,11 +52,11 @@ Installation using Fuel installer
Preparing the host to install Fuel by script
-.. Not all of these options are relevant for all scenarios. I advise following the
+.. Not all of these options are relevant for all scenarios. I advise following the
.. instructions applicable to the deploy tool used in the scenario.
-Before starting the installation of the os-odl-bgpnvp scenario some
-preparation of the machine that will host the Fuel VM must be done.
+Before starting the installation of the os-odl-bgpvpn-noha scenario the following
+preparation must be done on the machine that will host the Fuel VM.
Installation of required packages
@@ -64,17 +66,8 @@ Jumphost (or the host which serves the VMs for the virtual deployment) needs to
install the following packages:
- sudo apt-get install -y git make curl libvirt-bin libpq-dev qemu-kvm \
- qemu-system tightvncserver virt-manager sshpass \
- fuseiso genisoimage blackbox xterm python-pip \
- python-git python-dev python-oslo.config \
- python-pip python-dev libffi-dev libxml2-dev \
- libxslt1-dev libffi-dev libxml2-dev libxslt1-dev \
- expect curl python-netaddr p7zip-full
- sudo pip install GitPython pyyaml netaddr paramiko lxml scp \
- python-novaclient python-neutronclient python-glanceclient \
- python-keystoneclient debtcollector netifaces enum
+ sudo apt-get install -y git make curl libvirt-bin qemu-kvm \
+ python-pip python-dev
Download the source code and artifact
@@ -91,131 +84,32 @@ To check out a specific version of OPNFV, checkout the appropriate branch:
git checkout stable/gambia
-Now download the corresponding OPNFV Fuel ISO into an appropriate folder from
-the website https://www.opnfv.org/software/downloads/release-archives
-Have in mind that the fuel repo version needs to map with the downloaded
-artifact. Note: it is also possible to build the Fuel image using the
-tools found in the fuel git repository, but this is out of scope of the
-procedure described here. Check the Fuel project documentation for more
-information on building the Fuel ISO.
Simplified scenario deployment procedure using Fuel
-This section describes the installation of the os-odl-bgpvpn-ha or
+This section describes the installation of the
os-odl-bgpvpn-noha OPNFV reference platform stack across a server cluster
or a single host as a virtual deployment.
-dea.yaml and dha.yaml need to be copied and changed according to the lab-name/host
-where you deploy.
-Copy the full lab config from:
- cp <path-to-opnfv-fuel-repo>/deploy/config/labs/devel-pipeline/elx \
-Add at the bottom of dha.yaml
- fuel: 100G
- controller: 100G
- compute: 100G
- value: 4
- unit: KiB
- value: 16388608
- unit: KiB
- value: 16388608
-Check if the default settings in dea.yaml are in line with your intentions
-and make changes as required.
-We describe several alternative procedures in the following.
-First, we describe several methods that are based on the deploy.sh script,
-which is also used by the OPNFV CI system.
-It can be found in the Fuel repository.
-In addition, the SDNVPN feature can also be configured manually in the Fuel GUI.
-This is described in the last subsection.
-Before starting any of the following procedures, go to
- cd <opnfv-fuel-repo>/ci
-Full automatic virtual deployment High Availablity Mode
-The following command will deploy the high-availability flavor of SDNVPN scenario os-odl-bgpvpn-ha
-in a fully automatic way, i.e. all installation steps (Fuel server installation, configuration,
-node discovery and platform deployment) will take place without any further prompt for user input.
- sudo bash ./deploy.sh -b file://<path-to-opnfv-fuel-repo>/config/ -l devel-pipeline -p <your-lab-name> -s os-odl_l2-bgpvpn-ha -i file://<path-to-fuel-iso>
+This chapter describes how to deploy the scenario with the use of deploy.sh script,
+which is also used by the OPNFV CI system. Script can be found in the Fuel
Full automatic virtual deployment NO High Availability Mode
-The following command will deploy the SDNVPN scenario in its non-high-availability flavor (note the
-different scenario name for the -s switch). Otherwise it does the same as described above.
- sudo bash ./deploy.sh -b file://<path-to-opnfv-fuel-repo>/config/ -l devel-pipeline -p <your-lab-name> -s os-odl_l2-bgpvpn-noha -i file://<path-to-fuel-iso>
-Automatic Fuel installation and manual scenario deployment
-A useful alternative to the full automatic procedure is to only autodeploy the Fuel host and to run host selection, role assignment and SDNVPN scenario configuration manually.
+The following command will deploy the SDNVPN scenario in its non-high-availability flavor.
- sudo bash ./deploy.sh -b file://<path-to-opnfv-fuel-repo>/config/ -l devel-pipeline -p <your-lab-name> -s os-odl_l2-bgpvpn-ha -i file://<path-to-fuel-iso> -e
-With -e option the installer does not launch environment deployment, so
-a user can do some modification before the scenario is really deployed.
-Another interesting option is the -f option which deploys the scenario using an existing Fuel host.
-The result of this installation is a fuel sever with the right config for
-BGPVPN. Now the deploy button on fuel dashboard can be used to deploy the environment.
-It is as well possible to do the configuration manuell.
-Feature configuration on existing Fuel
-If a Fuel server is already provided but the fuel plugins for Opendaylight, Openvswitch
-and BGPVPN are not provided install them by:
- cd /opt/opnfv/
- fuel plugins --install fuel-plugin-ovs-*.noarch.rpm
- fuel plugins --install opendaylight-*.noarch.rpm
- fuel plugins --install bgpvpn-*.noarch.rpm
-If plugins are installed and you want to update them use --force flag.
-Now the feature can be configured. Create a new environment with "Neutron with ML2 plugin" and
-in there "Neutron with tunneling segmentation".
-Go to Networks/Settings/Other and check "Assign public network to all nodes". This is required for
-features such as floating IP, which require the Compute hosts to have public interfaces.
-Then go to settings/other and check "OpenDaylight plugin", "Use ODL to manage L3 traffic",
-"BGPVPN plugin" and set the OpenDaylight package version to "5.2.0-1". Then you should
-be able to check "BGPVPN extensions" in OpenDaylight plugin section.
-Now the deploy button on fuel dashboard can be used to deploy the environment.
+ ci/deploy.sh -l <lab_name> \
+ -p <pod_name> \
+ -b <URI to configuration repo containing the PDF file> \
+ -s os-odl-bgpvpn-noha \
+ -D \
+ -S <Storage directory for disk images> |& tee deploy.log
Virtual deployment using Apex installer