Fuel7.0安装

陈沙克在他的经典文章里分析了多个安装部署工具，用于openstack部署。

当然时代变了，所以很多这些工具都改头换面了。

目前看，

1 Mirantis Fuel仍然是最好的商业用选择，虽然还有瑕疵。

fuel在7和7之后已经提供开源的plugin插件，例如contrail的插件。并将fuel plugin方案放入openstack发展会中。

fuel在6的时候已经在提供商业化的plugin了。

fuel目前的8有个讨厌的地方就是需要联网取些内容（是我网的问题，联网去取docker image供slave用），对centos7.0的版本兼容性也不好。（是centos7.0的问题，其升kernel到7.1很大可能会挂）

fuel的瑕疵主要是有些场景会莫名的出现异常，部分删除和增加云会导致异常，一些硬盘大小识别在安装快结束时才告诉你容量不足。（主要是6的问题，8因网络问题至今无感）

2 RDO，即原来的packstack

RDO无光盘，也是出了名的网络安装，没网怎么活？

RDO安装的粒度必须是全openstack

RDO安装了ceilemeter，第二次换了ip（如果是dhcp申请的），重装，如果不选择ceilemeter，则原来的ceil不会自动消亡，还在那里盲目发信息要求通讯。

3 其他puppet类，如陈所言。

其中rackspace，我没找到光盘，也没找到脚本。rackspace是不是已经不怎么玩openstack了？

4 ansible和kolla

ansible比较方便，因为总有人讨厌puppet，在下不懂puppet，所以也不愿意碰

kolla是ansible语言的上层开发。提供了以openstack 服务级粒度的openstack安装部署。

kolla可以支持对失败的openstack的nova，keystone服务重新单独部署。

kolla还基于docker

fuel7.0安装，和fuel5.0、6安装相似，这次小弟是在virtualbox上安装的。

小弟修改了config.sh

主要改动点

1 master， 1compute, 1controller,

给计算节点多点资源。

不过安装成功后，惊讶的发现fuel7很智能的把整张盘的大小都给了compute，controller

内存上各节点的配置则和config.sh相同。

贴上config.sh

#!/bin/bash#    Copyright 2013 Mirantis, Inc.##    Licensed under the Apache License, Version 2.0 (the "License"); you may#    not use this file except in compliance with the License. You may obtain#    a copy of the License at##         http://www.apache.org/licenses/LICENSE-2.0##    Unless required by applicable law or agreed to in writing, software#    distributed under the License is distributed on an "AS IS" BASIS, WITHOUT#    WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the#    License for the specific language governing permissions and limitations#    under the License.source ./functions/memory.sh# Get the first available ISO from the directory 'iso'iso_path=`ls -1t iso/*.iso 2>/dev/null | head -1`# Every Mirantis OpenStack machine name will start from this prefixvm_name_prefix=fuel-# By default, all available network interfaces vboxnet won't be removed,# if their IP addresses don't match with fuel_master_ips (10.20.0.1 172.16.0.254# 172.16.1.1)# If you want to remove all existing vbox interfaces, then use rm_network=1# 0 - don't remove all vbox networks. Remove only fuel networks if they exist# 1 - remove all vbox networksrm_network=0# Please add the IPs accordingly if you going to create non-default NICs number# 10.20.0.1/24   - Mirantis OpenStack Admin network# 172.16.0.1/24  - OpenStack Public/External/Floating network# 172.16.1.1/24  - OpenStack Fixed/Internal/Private network# 192.168.0.1/24 - OpenStack Management network# 192.168.1.1/24 - OpenStack Storage network (for Ceph, Swift etc)fuel_master_ips="10.20.0.1 172.16.0.254 172.16.1.1"# Network mask for fuel interfacesmask="255.255.255.0"# Determining the type of operating system and adding CPU core to the master node  case "$(uname)" in    Linux)      os_type="linux"      if [ "$(nproc)" -gt "1" ]; then        vm_master_cpu_cores=2      else        vm_master_cpu_cores=1      fi    ;;    Darwin)      os_type="darwin"      mac_nproc=`sysctl -a | grep machdep.cpu.thread_count | sed 's/^machdep.cpu.thread_count\:[ \t]*//'`      if [ "$mac_nproc" -gt "1" ]; then        vm_master_cpu_cores=2      else        vm_master_cpu_cores=1      fi    ;;    CYGWIN*)      os_type="cygwin"      if [ "$(nproc)" -gt "1" ]; then        vm_master_cpu_cores=2      else        vm_master_cpu_cores=1      fi    ;;    *)      echo "$(uname) is not supported operating system."      exit 1    ;;  esac# Master node settingsvm_master_memory_mb=1536vm_master_disk_mb=65535# Master node access to the internet through the host system, using VirtualBox NAT adaptervm_master_nat_network=192.168.200.0/24vm_master_nat_gateway=192.168.200.2# These settings will be used to check if master node has installed or not.# If you modify networking params for master node during the boot time#   (i.e. if you pressed Tab in a boot loader and modified params),#   make sure that these values reflect that change.vm_master_ip=10.20.0.2vm_master_username=rootvm_master_password=r00tmevm_master_prompt='root@fuel ~]#'# The number of nodes for installing OpenStack on#   - for minimal non-HA installation, specify 2 (1 controller + 1 compute)#   - for minimal non-HA with Cinder installation, specify 3 (1 ctrl + 1 compute + 1 cinder)#   - for minimal HA installation, specify 4 (3 controllers + 1 compute)if [ "$CONFIG_FOR" = "16GB" ]; then  cluster_size=5elif [ "$CONFIG_FOR" = "8GB" ]; then  cluster_size=3else  # Section for custom configuration  cluster_size=2fi# Slave node settings. This section allows you to define CPU count for each slave node.# You can specify CPU count for your nodes as you wish, but keep in mind resources of your machine.# If you don't, then will be used default parameterif [ "$CONFIG_FOR" = "16GB" ]; then  vm_slave_cpu_default=1  vm_slave_cpu[1]=1  vm_slave_cpu[2]=1  vm_slave_cpu[3]=1  vm_slave_cpu[4]=1  vm_slave_cpu[5]=1elif [ "$CONFIG_FOR" = "8GB" ]; then  vm_slave_cpu_default=1  vm_slave_cpu[1]=1  vm_slave_cpu[2]=1  vm_slave_cpu[3]=1else  # Section for custom configuration  vm_slave_cpu_default=1  vm_slave_cpu[1]=4  vm_slave_cpu[2]=1fi# This section allows you to define RAM size in MB for each slave node.# Keep in mind that PXE boot might not work correctly with values lower than 768.# You can specify memory size for the specific slaves, other will get default vm_slave_memory_default# Mirantis OpenStack 3.2 controllers require 1280 MiB of RAM as absolute minimum due to Heat!# You may comment out all the following memory parameters to use default value for each node.# It is recommended if you going to try HA configurations.# for controller node at least 1.5Gb is required if you also run Ceph and Heat on it# and for Ubuntu controller we need 2Gb of ram# For compute node 1GB is recommended, otherwise VM instances in OpenStack may not boot# For dedicated Cinder, 768Mb is OK, but Ceph needs 1Gb minimumif [ "$CONFIG_FOR" = "16GB" ]; then  vm_slave_memory_default=1536  vm_slave_memory_mb[1]=2048  vm_slave_memory_mb[2]=2048  vm_slave_memory_mb[3]=2048  vm_slave_memory_mb[4]=2048  vm_slave_memory_mb[5]=2048elif [ "$CONFIG_FOR" = "8GB" ]; then  vm_slave_memory_default=1024  vm_slave_memory_mb[1]=1536  vm_slave_memory_mb[2]=1536  vm_slave_memory_mb[3]=1536else  # Section for custom configuration  vm_slave_memory_default=1024  vm_slave_memory_mb[1]=12048  vm_slave_memory_mb[2]=1024  vm_slave_memory_mb[3]=1024fi# Within demo cluster created by this script, all slaves (controller# and compute nodes) will have identical disk configuration. Each# slave will have three disks with sizes defined by the variables below. In a disk configuration# dialog you will be able to allocate the whole disk or it's part for# operating system (Base OS), VMs (Virtual Storage), Ceph or other function,# depending on the roles applied to the server.# Nodes with combined roles may require more disk space.vm_slave_first_disk_mb=85535vm_slave_second_disk_mb=35535vm_slave_third_disk_mb=65535# Set to 1 to run VirtualBox in headless modeheadless=0skipfuelmenu="no"