first draft of 3 installers for Auto

JIRA: AUTO-42
JIRA: AUTO-43
JIRA: AUTO-46
JIRA: AUTO-50

a) Moved 3 scripts (Fuel/MCP, Compass4nfv, Apex/TripleO) from
Auto Wiki pages to Auto repository. Did some cross-script alignment.
Host folder in repository is: auto/ci. Abandoned JOID script:
JOID seems to be unsupported after Fraser.

b) Further fixes to these scripts can now be done in the Auto
repository (not in Wiki), and inclusion in Auto CI jobs can also be
tested.

c) File names might be revisited, to account for variants: CPU, OS,
scenario (openstack/kubernetes, HA/NOHA), size configs (vCPUs, RAM),
etc. Or explore having these parameters as script parameters, maybe
collapsing all scripts into one, where the installer would also be a
script parameter.

d) This batch of scripts is for x86 architectures; future
versions/additions will cover Arm architectures;

e1) Fuel/MCP script works (Fuel@x86)

e2) Compass script works (x86, virtual deploy, NOHA)

e3) Apex works on 2nd attempt; first attempt still has issues on
blank servers (due to undocumented initial setup steps, causing
errors which don't happen on servers where several tools and
packages have already been installed).
raised ticket on Apex: https://jira.opnfv.org/browse/APEX-633

f) scripts all pass the shellcheck tool

Change-Id: Ib9a393bcc6f94ec8be82d5d2d4c9c194a018c0f7
Signed-off-by: Gerard Damm <gerard.damm@wipro.com>

1 files changed, 209 insertions, 0 deletions

diff --git a/ci/deploy-opnfv-apex-centos.sh b/ci/deploy-opnfv-apex-centos.sh
new file mode 100644
index 0000000..a3a0433
--- /dev/null
+++ b/ci/deploy-opnfv-apex-centos.sh
@@ -0,0 +1,209 @@
+#!/usr/bin/env bash
+
+# /usr/bin/env bash or /bin/bash ? /usr/bin/env bash is more environment-independent
+# beware of files which were edited in Windows, and have invisible \r end-of-line characters, causing Linux errors
+
+##############################################################################
+# Copyright (c) 2018 Wipro Limited and others.
+#
+# All rights reserved. This program and the accompanying materials
+# are made available under the terms of the Apache License, Version 2.0
+# which accompanies this distribution, and is available at
+# http://www.apache.org/licenses/LICENSE-2.0
+##############################################################################
+
+# OPNFV contribution guidelines Wiki page:
+# https://wiki.opnfv.org/display/DEV/Contribution+Guidelines
+
+# OPNFV/Auto project:
+# https://wiki.opnfv.org/pages/viewpage.action?pageId=12389095
+
+
+# localization control: force script to use default language for output, and force sorting to be bytewise
+# ("C" is from C language, represents "safe" locale everywhere)
+# (result: the script will consider only basic ASCII characters and disable UTF-8 multibyte match)
+export LANG=C
+export LC_ALL=C
+
+##################################################################################
+## installation of OpenStack via OPNFV Apex/TripleO, on CentOS, virtual deployment
+##################################################################################
+# reference manual: https://docs.opnfv.org/en/latest/submodules/apex/docs/release/installation/index.html
+# page for virtual deployment: https://docs.opnfv.org/en/latest/submodules/apex/docs/release/installation/virtual.html
+
+echo "*** begin AUTO install: OPNFV Apex/TripleO"
+
+# check OS version
+echo "*** print OS version (must be CentOS, version 7 or more)"
+cat /etc/*release
+
+# Manage Nested Virtualization
+echo "*** ensure Nested Virtualization is enabled on Intel x86"
+echo "***   nested flag before:"
+cat /sys/module/kvm_intel/parameters/nested
+rm -f /etc/modprobe.d/kvm-nested.conf
+{ printf "options kvm-intel nested=1\n";\
+  printf "options kvm-intel enable_shadow_vmcs=1\n";\
+  printf "options kvm-intel enable_apicv=1\n";\
+  printf "options kvm-intel ept=1\n"; } >> /etc/modprobe.d/kvm-nested.conf
+sudo modprobe -r kvm_intel
+sudo modprobe -a kvm_intel
+echo "***   nested flag after:"
+cat /sys/module/kvm_intel/parameters/nested
+
+echo "*** verify status of modules in the Linux Kernel: kvm_intel module should be loaded for x86_64 machines"
+lsmod | grep kvm_
+grep kvm_ < /proc/modules
+
+# 3 additional pre-installation preparations, lifted from OPNFV/storperf (they are post-installation there):
+# https://wiki.opnfv.org/display/storperf/LaaS+Setup+For+Development#LaaSSetupForDevelopment-InstallOPNFVApex
+# (may of may not be needed, to enable first-time Apex installation on blank server)
+
+# 1) Install Docker
+sudo yum install -y yum-utils device-mapper-persistent-data lvm2
+sudo yum-config-manager --add-repo https://download.docker.com/linux/centos/docker-ce.repo
+sudo yum install -y docker-ce
+sudo systemctl start docker
+
+# 2) Install docker-compose
+sudo curl -L "https://github.com/docker/compose/releases/download/1.21.2/docker-compose-$(uname -s)-$(uname -m)" -o /usr/local/bin/docker-compose
+sudo chmod +x /usr/local/bin/docker-compose
+
+# 3) Install Python
+sudo yum install -y python-virtualenv
+sudo yum groupinstall -y "Development Tools"
+sudo yum install -y openssl-devel
+
+
+# update everything (upgrade: riskier than update, as packages supposed to be unused will be deleted)
+# (note: can take several minutes; may not be necessary)
+sudo yum -y update
+
+
+# download Apex packages
+echo "*** downloading packages:"
+sudo yum -y install https://repos.fedorapeople.org/repos/openstack/openstack-pike/rdo-release-pike-1.noarch.rpm
+sudo yum -y install epel-release
+# note: EPEL = Extra Packages for Enterprise Linux
+sudo curl -o /etc/yum.repos.d/opnfv-apex.repo http://artifacts.opnfv.org/apex/fraser/opnfv-apex.repo
+
+# install three required RPMs (RedHat/RPM Package Managers); this takes several minutes
+sudo yum -y install http://artifacts.opnfv.org/apex/fraser/opnfv-apex-6.2.noarch.rpm http://artifacts.opnfv.org/apex/fraser/opnfv-apex-undercloud-6.2.noarch.rpm http://artifacts.opnfv.org/apex/fraser/opnfv-apex-python34-6.2.noarch.rpm
+
+# clean-up old Apex versions if any
+## precautionary opnfv-clean doesn't work... (even though packages are installed at this point)
+opnfv-clean
+
+# Manage DNS references
+# probably not needed on an already configured server: already has DNS references
+# echo "nameserver 8.8.8.8" >> /etc/resolv.conf
+echo "*** printout of /etc/resolv.conf :"
+cat /etc/resolv.conf
+
+# prepare installation directory
+mkdir -p /opt/opnfv-TripleO-apex
+cd /opt/opnfv-TripleO-apex
+
+# make sure cp is not aliased or a function; same for mv and rm
+unalias cp
+unset -f cp
+unalias mv
+unset -f mv
+unalias rm
+unset -f rm
+
+# 2 YAML files from /etc/opnfv-apex/ are needed for virtual deploys:
+# 1) network_settings.yaml : may need to update NIC names, to match the NIC names on the deployment server
+# 2) standard scenario file (os-nosdn-nofeature-noha.yaml, etc.), or customized deploy_settings.yaml
+
+# make a local copy of YAML files (not necessary: could deploy from /etc/opnfv-apex); local copies are just for clarity
+# 1) network settings
+cp /etc/opnfv-apex/network_settings.yaml .
+# 2) deploy settings
+# copy one of the 40+ pre-defined scenarios (one of the YAML files)
+# for extra customization, git clone Apex repo, and copy and customize the generic deploy_settings.yaml
+# git clone https://git.opnfv.org/apex
+# cp ./apex/config/deploy/deploy_settings.yaml .
+cp /etc/opnfv-apex/os-nosdn-nofeature-noha.yaml ./deploy_settings.yaml
+# cp /etc/opnfv-apex/os-nosdn-nofeature-ha.yaml ./deploy_settings.yaml
+
+# Note: content of os-nosdn-nofeature-noha.yaml
+# ---
+# global_params:
+#   ha_enabled: false
+#
+# deploy_options:
+#   sdn_controller: false
+#   tacker: true
+#  congress: true
+#   sfc: false
+#   vpn: false
+
+
+# modify NIC names in network settings YAML file, specific to your environment (e.g. replace em1 with ens4f0 in LaaS)
+# Note: actually, this should not matter for a virtual environment
+sed -i 's/em1/ens4f0/' network_settings.yaml
+
+# launch deploy (works if openvswitch module is installed, which may not be the case the first time around)
+echo "*** deploying OPNFV by TripleO/Apex:"
+# --debug for detailed debug info
+# -v: Enable virtual deployment
+# note: needs at least 10G RAM for controllers
+sudo opnfv-deploy --debug -v -n network_settings.yaml -d deploy_settings.yaml
+# without --debug:
+# sudo opnfv-deploy -v -n network_settings.yaml -d deploy_settings.yaml
+
+# with specific sizing:
+# sudo opnfv-deploy --debug -v -n network_settings.yaml -d deploy_settings.yaml --virtual-compute-ram 32 --virtual-cpus 16 --virtual-computes 4
+
+
+# verify that the openvswitch module is listed:
+lsmod | grep openvswitch
+grep openvswitch < /proc/modules
+
+##{
+## workaround: do 2 successive installations... not exactly optimal...
+## clean up, as now opnfv-clean should work
+#opnfv-clean
+## second deploy try, should succeed (whether first one failed or succeeded)
+#sudo opnfv-deploy -v -n network_settings.yaml -d deploy_settings.yaml
+##}
+
+
+
+# verifications: https://docs.opnfv.org/en/latest/submodules/apex/docs/release/installation/verification.html
+
+# {
+# if error after deploy.sh: "libvirt.libvirtError: Storage pool not found: no storage pool with matching name 'default'"
+
+# This usually happens if for some reason you are missing a default pool in libvirt:
+# $ virsh pool-list |grep default
+# You can recreate it manually:
+# $ virsh pool-define-as default dir --target /var/lib/libvirt/images/
+# $ virsh pool-autostart default
+# $ virsh pool-start default
+# }
+
+# {
+# if error after deploy.sh: iptc.ip4tc.IPTCError
+# check Apex jira ticket #521 https://jira.opnfv.org/browse/APEX-521
+# }
+
+# OpenvSwitch should not be missing, as it is a requirement from the RPM package:
+# https://github.com/opnfv/apex/blob/stable/fraser/build/rpm_specs/opnfv-apex-common.spec#L15
+
+
+
+# install python 3 on CentOS
+echo "*** begin install python 3.6 (3.4 should be already installed by default)"
+
+sudo yum -y install python36
+# install pip and setup tools
+sudo curl -O https://bootstrap.pypa.io/get-pip.py
+hash -r
+sudo /usr/bin/python3.6 get-pip.py --no-warn-script-location
+
+
+
+echo "*** end AUTO install: OPNFV Apex/TripleO"
+