前言
SmartX专注于 IT 基础架构创新，提供全球领先的分布式存储、计算引擎、网络等 IT 基础设施产品，打造最硬核、最稳定的企业云、分布式存储和云原生产品解决方案。
其主打软件SMTX OS是SmartX 超融合基础架构核心软件， 内置了分布式块存储 ZBS、原生虚拟化 ELF 以及异步复制、同城双活、网络与安全等高级组件。支持 VMware 虚拟化融合部署。
据官网介绍，在使用方式上SMTX OS可以单独安装在主机上提供基于KVM的计算虚拟化（面前免费）和分布式块存储ZBS；也可以和例如VMware的vSphere计算虚拟化产品一起组成超融合系统。
本实验为第一种情况，实践在主机上安装SMTX OS并实现计算和存储虚拟化。
实验环境及参数
设备情况表
每台主机两块960G的SSD作为Cache每台主机三块2.3T的盘作为数据盘实验拓扑图
网关192.168.80.1/20，同时提供DHCP服务两台物理交换机分别提供1G和10G网络接入设备管理和业务共享一个千兆出口主机内部架构与连接
IP地址设置
Install 3.5.16版本
此版本比较老，用于模拟老用户的环境
Install OS
使用iDRAC口登录并启动虚拟控制台
连接虚拟介质，映射到本地iso文件
选择启动→开机控制→虚拟CD启动
重启设备进入下面界面，键入回车后自动安装
如果安装过程中系统确定不了安装的盘，需要人工键入（这里键入sdf）：
每台设备都需要完成以上1~4步骤
SMTX OS集群初始化
安装完成以后会出现login CLI，初始化用户名和密码为：root/HC!r0cks
由于环境提供了DHCP功能，此时可以在任意一台看到分配的IP地址；如果没有地址，配置方法和CentOS一致。
使用浏览器http://192.168.91.178
按提示进入集群初始化
a. 键入集群名字
b. 选择计算虚拟化平台
在这里也可以看到，该软件和可以不同的虚拟化平台结合，如vSphere和XenServer
系统会扫描（L2层）集群可用的设备，这个时候可以改主机的名字
配置集群网络
a. 初始化需要配置两台分布式交换机
b. 一台是业务或管理网络交换机
c. 一台是存储网络交换机，一定要选择>10G的物理网络端口
按照设计，填入IP地址
a. 如果选择的端口和主机管理关口一致，这些地址会替代DHCP得到的IP地址，也可以事先在主机上配置好
b. 配置DNS和NTP
检查配置以后就可以执行集群部署了
完成以后，点击开始部署后配置，系统转到集群配置的第一个IP地址的web页面
配置管理员密码和IPMI地址
配置好了以后再次登录会进入到管理界面，名字叫fisheye
Upgrade to 4.0.9
在3.6.15这个不版本中只支持命令行升级，以下是起升级的步骤。
选择任意一台主机，传入升级使用的ISO文件（scp或工具）建议ssh到相应的主机 把iso文件做成Yum源
mkdir /mnt/isomount -o loop SMTXOS-4.0.9-el7-2101291745-x86_64.iso /mnt/isomkdir /etc/yum.repos.d/bkmv /etc/yum.repos.d/*.repo /etc/yum.repos.d/bk/cat > /etc/yum.repos.d/smartxos.repo << EOF[smartxos-local-iso]name=smartxosbaseurl=file:///mnt/isogpgcheck=0enabled=1EOF
更新对应的 smartx-upgrade rpm包
yum clean allyum update smartx-upgrade -y
[root@smtx-01 17:47:59 ~]$yum clean allyum update smartx-upgrade -yLoaded plugins: fastestmirror, langpacksCleaning repos: smartxos-local-isoCleaning up everythingMaybe you want: rm -rf /var/cache/yum, to also free up space taken by orphaned data from disabled or removed repos[root@smtx-01 17:48:07 ~]$yum update smartx-upgrade -yLoaded plugins: fastestmirror, langpackssmartxos-local-iso | 3.6 kB 00:00:00(1/2): smartxos-local-iso/group_gz | 156 kB 00:00:00(2/2): smartxos-local-iso/primary_db| 1.3 MB 00:00:00Determining fastest mirrorsResolving Dependencies--> Running transaction check---> Package smartx-upgrade.x86_64 0:3.5.16-rc5.0.git.7eca0010e.el7.centos will be updated---> Package smartx-upgrade.x86_64 0:4.0.9-rc20.0.git.g0aed4c1bf.el7.centos will be an update--> Finished Dependency ResolutionDependencies Resolved=========================================================================================================================================================================Package Arch Version RepositorySize=========================================================================================================================================================================Updating:smartx-upgrade x86_64 4.0.9-rc20.0.git.g0aed4c1bf.el7.centos smartxos-local-iso 71 kTransaction Summary=========================================================================================================================================================================Upgrade 1 PackageTotal download size: 71 kDownloading packages:Running transaction checkRunning transaction testTransaction test succeededRunning transactionUpdating : smartx-upgrade-4.0.9-rc20.0.git.g0aed4c1bf.el7.centos.x86_641/2 Cleanup : smartx-upgrade-3.5.16-rc5.0.git.7eca0010e.el7.centos.x86_64 2/2 Verifying : smartx-upgrade-4.0.9-rc20.0.git.g0aed4c1bf.el7.centos.x86_641/2 Verifying : smartx-upgrade-3.5.16-rc5.0.git.7eca0010e.el7.centos.x86_64 2/2 Updated:smartx-upgrade.x86_64 0:4.0.9-rc20.0.git.g0aed4c1bf.el7.centosComplete!
进⼊ /usr/share/upgrade/runner/ ⽬录，执⾏以下脚本。其中–iso_path跟iso文件的绝对目录
cd /usr/share/upgrade/runner/nohup python cluster_upgrader.py --iso_path /root/SMTXOS-4.0.9-el7-2101291745-x86_64.iso &[root@smtx-01 17:48:11 ~]$cd /usr/share/upgrade/runner/[root@smtx-01 17:51:30 runner]$nohup python cluster_upgrader.py --iso_path /root/SMTXOS-4.0.9-el7-2101291745-x86_64.iso &[1] 90862nohup: ignoring input and appending output to ‘nohup.out’
系统没有更多提示，可以使用tailf nohup.out 实时查看更新日志
[root@smtx-01 17:51:49 runner]$tailf nohup.out -04-08 17:51:42,918 INFO run: 10.10.87.218:2181 follower0x1000004f4-04-08 17:51:44,019 INFO _connect: Connecting to 10.10.87.216:2181-04-08 17:51:44,040 INFO _session_callback: Zookeeper connection established, state: CONNECTED-04-08 17:51:44,057 INFO _connect_attempt: Closing connection to 10.10.87.216:2181-04-08 17:51:44,057 INFO _session_callback: Zookeeper session lost, state: CLOSED-04-08 17:51:44,058 INFO upgrade_cluster: pre_upgrade_cmd: cd /usr/share/upgrade/playbook; ansible-playbook -i /usr/share/upgrade/playbook/inventory cluster_pre_upgrade.yaml --extra-vars '{"iso_file_name": "SMTXOS-4.0.9-el7-2101291745-x86_64.iso", "iso_file_path": "/root/SMTXOS-4.0.9-el7-2101291745-x86_64.iso"}'-04-08 17:51:45,388 INFO run: -04-08 17:51:45,388 INFO run: PLAY [localhost] ***************************************************************-04-08 17:51:45,419 INFO run: -04-08 17:51:45,420 INFO run: TASK [check storage status, without dead pextent or recover pextent] ***********-04-08 17:53:58,504 INFO run: changed: [localhost]-04-08 17:53:58,505 INFO run: -04-08 17:53:58,505 INFO run: PLAY [cluster] *****************************************************************-04-08 17:53:58,505 INFO run: -04-08 17:53:58,505 INFO run: TASK [umount iso mount path if exist] ******************************************-04-08 17:53:59,042 INFO run: fatal: [10.10.87.217]: FAILED! => {"changed": true, "cmd": ["umount", "/mnt/iso"], "delta": "0:00:00.011607", "end": "-04-08 17:53:58.994951", "msg": "non-zero return code", "rc": 32, "start": "-04-08 17:53:58.983344", "stderr": "umount: /mnt/iso: mountpoint not found", "stderr_lines": ["umount: /mnt/iso: mountpoint not found"], "stdout": "", "stdout_lines": []}-04-08 17:53:59,042 INFO run: ...ignoring-04-08 17:53:59,051 INFO run: changed: [10.10.87.216]-04-08 17:53:59,084 INFO run: fatal: [10.10.87.218]: FAILED! => {"changed": true, "cmd": ["umount", "/mnt/iso"], "delta": "0:00:00.009553", "end": "-04-08 17:53:59.029823", "msg": "non-zero return code", "rc": 32, "start": "-04-08 17:53:59.020270", "stderr": "umount: /mnt/iso: mountpoint not found", "stderr_lines": ["umount: /mnt/iso: mountpoint not found"], "stdout": "", "stdout_lines": []}-04-08 17:53:59,084 INFO run: ...ignoring-04-08 17:53:59,084 INFO run: -04-08 17:53:59,084 INFO run: TASK [check update_upgrade.progress.py exists] *********************************-04-08 17:53:59,485 INFO run: ok: [10.10.87.217]-04-08 17:53:59,488 INFO run: ok: [10.10.87.218]-04-08 17:53:59,503 INFO run: ok: [10.10.87.216]-04-08 17:53:59,504 INFO run: -04-08 17:53:59,504 INFO run: TASK [update node upgrade status] **********************************************-04-08 17:53:59,543 INFO run: skipping: [10.10.87.218]-04-08 17:53:59,544 INFO run: skipping: [10.10.87.217]-04-08 17:54:00,454 INFO run: changed: [10.10.87.216]-04-08 17:54:00,455 INFO run: -04-08 17:54:00,455 INFO run: TASK [remove all repo files] ***************************************************-04-08 17:54:00,787 INFO run: changed: [10.10.87.216]-04-08 17:54:00,794 INFO run: changed: [10.10.87.217]-04-08 17:54:00,795 INFO run: changed: [10.10.87.218]-04-08 17:54:00,806 INFO run: -04-08 17:54:00,806 INFO run: TASK [copy iso file] ***********************************************************......-04-08 18:16:07,877 INFO run: PLAY [localhost] ***************************************************************-04-08 18:16:07,877 INFO run: -04-08 18:16:07,877 INFO run: TASK [systemd] *****************************************************************-04-08 18:16:08,647 INFO run: changed: [localhost]-04-08 18:16:08,648 INFO run: -04-08 18:16:08,648 INFO run: PLAY RECAP *********************************************************************-04-08 18:16:08,648 INFO run: 10.10.87.217: ok=1 changed=1 unreachable=0 failed=0-04-08 18:16:08,648 INFO run: 10.10.87.218: ok=1 changed=1 unreachable=0 failed=0-04-08 18:16:08,648 INFO run: localhost : ok=1 changed=1 unreachable=0 failed=0-04-08 18:16:08,648 INFO run: -04-08 18:16:09,748 INFO upgrade_cluster: Cluster upgrade successful
整个过程有十几分钟，看到 upgrade_cluster: Cluster upgrade successful，代表集群升级成功
在fisheys上查看
Upgrade to 4.0.10
SMTX OS 4.0.10是现役使用的主力版本，在我们的实验中4.0.9可以支持web方式的升级。
点击设置
在软件升级中点击管理
上传相应的iso文件和json文件
点击升级成功以后在首页上可以看到现在的版本信息
以上
