通过Azure File Service搭建基于iscsi的共享盘

在Azure上目前已经有基于Samba协议的共享存储了。

但目前在Azure上，还不能把Disk作为共享盘。而在实际的应用部署中，共享盘是做集群的重要组件之一。比如仲裁盘、Shared Disk等。

本文将介绍，如果通过基于Samba的文件共享，加上Linux的Target、iscsid以及multipath等工具，提供带HA的共享Disk。

上图是整体架构：

通过两台CentOS7.2的VM都挂载Azure File；在Azure File上创建一个文件: disk.img，两个VM通过iscsi server的软件target，同时把这个disk.img作为iscsi的disk发布出去；一台装有iscsid的CentOS7.2的Server同时挂载这两个iscsi Disk；再采用multipath的软件把这两个盘合成一个。

在这种架构下，iscsi客户端获得了一块iscsi的disk。而且这块Disk是提供HA的网络Disk。

具体实现方式如下：

一、 创建File Service

1. 在Azure Portal上创建File Service

在Storage Account中点击Add：

填写相关信息，点击Create。

创建成功后，在Storage Account中选择创建好的Storage Account：

选择File：

点击+file Share后，填写File Share的名字，点击创建。

创建好后，可以看到在Linux中mount的命令提示：

在Access Keys中复制key

2. 在两台iscsi server上mount这个File Service

首先查看服务器版本：

[root@hwis01 ~]# cat /etc/redhat-release
CentOS Linux release 7.2. (Core)

都是CentOS7.0以上的版本，可以支持Samba3.0。

根据前面File Service的信息执行下面的命令：

[root@hwis01 ~]# mkdir /file
[root@hwis01 ~]# sudo mount -t cifs //hwiscsi.file.core.chinacloudapi.cn/hwfile /file -o vers=3.0,username=hwiscsi,password=xxxxxxxx==,dir_mode=0777,file_mode=0777
[root@hwis01 ~]# df -h
Filesystem Size Used Avail Use% Mounted on
/dev/sda1 30G .1G 29G % /
devtmpfs 829M  829M % /dev
tmpfs 839M  839M % /dev/shm
tmpfs 839M 8.3M 831M % /run
tmpfs 839M  839M % /sys/fs/cgroup
/dev/sdb1 69G 53M 66G % /mnt/resource
tmpfs 168M  168M % /run/user/
//hwiscsi.file.core.chinacloudapi.cn/hwfile 5.0T 0 5.0T 0% /file

二、 在iscsi server上创建iscsi的disk

1. 在共享目录中创建disk.img

[root@hwis01 ~]# dd if=/dev/zero of=/file/disk.img bs=1M count=
+ records in
+ records out
 bytes (1.1 GB) copied, 20.8512 s, 51.5 MB/s

本机查看

[root@hwis01 ~]# cd /file
[root@hwis01 file]# ll
total
-rwxrwxrwx.  root root  Nov  : disk.img

在另外一台Server上查看：

[root@hwis02 ~]# cd /file
[root@hwis02 file]# ll
total
-rwxrwxrwx.  root root  Nov  : disk.img

2.安装相关软件

iscsi服务器端安装targetcli：

[root@hwis01 file]# yum install -y targetcli

iscsi客户端安装iscsi-initiator-utils:

[root@hwic01 ~]# yum install iscsi-initiator-utils -y

安装完成后，在iscsi的客户端机器上查看iqn号码：

[root@hwic01 /]# cd /etc/iscsi/
[root@hwic01 iscsi]# more initiatorname.iscsi
InitiatorName=iqn.-.hw.ic01:client

3.用targetcli创建iscsi的disk

[root@hwis01 file]# targetcli
Warning: Could not load preferences file /root/.targetcli/prefs.bin.
targetcli shell version 2.1.fb41
Copyright - by Datera, Inc and others.
For help on commands, type 'help'.
/> ls
o- / ....................................................................................................................... [...]
o- backstores .............................................................................................................. [...]
| o- block .................................................................................................. [Storage Objects: ]
| o- fileio ................................................................................................. [Storage Objects: ]
| o- pscsi .................................................................................................. [Storage Objects: ]
| o- ramdisk ................................................................................................ [Storage Objects: ]
o- iscsi ............................................................................................................ [Targets: ]
o- loopback ......................................................................................................... [Targets: ]
/> cd backstores/
/backstores> cd fileio
/backstores/fileio> create disk01 /file/disk.img 1G
/file/disk.img exists, using its size ( bytes) instead
Created fileio disk01 with size
/backstores/fileio> cd /iscsi/
/iscsi> create iqn.-.hw.is01:disk01.lun0
Created target iqn.-.hw.is01:disk01.lun0.
Created TPG .
Global pref auto_add_default_portal=true
Created default portal listening on all IPs (0.0.0.0), port .
/iscsi> cd iqn.-.hw.is01:disk01.lun0/tpg1/luns/
/iscsi/iqn....un0/tpg1/luns> create /backstores/fileio/disk01
Created LUN .
/iscsi/iqn....un0/tpg1/luns> cd ../acls/
/iscsi/iqn....un0/tpg1/acls> create iqn.-.hw.ic01:client
Created Node ACL for iqn.-.hw.ic01:client
Created mapped LUN .
/iscsi/iqn....un0/tpg1/acls> ls
o- acls ................................................................................................................ [ACLs: ]
o- iqn.-.hw.ic01:client ................................................................................... [Mapped LUNs: ]
o- mapped_lun0 ......................................................................................... [lun0 fileio/disk01 (rw)]
/iscsi/iqn....un0/tpg1/acls> cd /
/> ls
o- / ....................................................................................................................... [...]
o- backstores .............................................................................................................. [...]
| o- block .................................................................................................. [Storage Objects: ]
| o- fileio ................................................................................................. [Storage Objects: ]
| | o- disk01 ..................................................................... [/file/disk.img (.0GiB) write-back activated]
| o- pscsi .................................................................................................. [Storage Objects: ]
| o- ramdisk ................................................................................................ [Storage Objects: ]
o- iscsi ............................................................................................................ [Targets: ]
| o- iqn.-.hw.is01:disk01.lun0 ................................................................................... [TPGs: ]
| o- tpg1 ................................................................................................. [no-gen-acls, no-auth]
| o- acls .............................................................................................................. [ACLs: ]
| | o- iqn.-.hw.ic01:client ............................................................................... [Mapped LUNs: ]
| | o- mapped_lun0 ..................................................................................... [lun0 fileio/disk01 (rw)]
| o- luns .............................................................................................................. [LUNs: ]
| | o- lun0 ..................................................................................... [fileio/disk01 (/file/disk.img)]
| o- portals ........................................................................................................ [Portals: ]
| o- 0.0.0.0: ........................................................................................................... [OK]
o- loopback ......................................................................................................... [Targets: ]

查看配置文件中的wwn

[root@hwis01 file]# cd /etc/target
[root@hwis01 target]# ls
backup saveconfig.json
[root@hwis01 target]# vim saveconfig.json
[root@hwis01 target]# grep wwn saveconfig.json
"wwn": "acadb3f7-9a2d-44f4-8caf-de627ea98e27"
"node_wwn": "iqn.2016-10.hw.ic01:client"
"wwn": "iqn.2016-10.hw.is01:disk01.lun0"

记录下第一个disk的 "wwn": "acadb3f7-9a2d-44f4-8caf-de627ea98e27"

并将其复制到iscsi server2中的配置。

在server2上查看相关信息：

[root@hwis02 target]# grep wwn saveconfig.json
"wwn": "acadb3f7-9a2d-44f4-8caf-de627ea98e27"
"node_wwn": "iqn.2016-10.hw.ic01:client"
"wwn": "iqn.2016-10.hw.is02:disk01.lun0"

4.开启target服务

[root@hwis01 target]# systemctl enable target
Created symlink from /etc/systemd/system/multi-user.target.wants/target.service to /usr/lib/systemd/system/target.service.
[root@hwis01 target]# systemctl start target
[root@hwis01 target]# systemctl status target
target.service - Restore LIO kernel target configuration
Loaded: loaded (/usr/lib/systemd/system/target.service; enabled; vendor preset: disabled)
Active: active (exited) since Fri -- :: UTC; 7s ago
Process:  ExecStart=/usr/bin/targetctl restore (code=exited, status=/SUCCESS)
Main PID:  (code=exited, status=/SUCCESS)
Nov  :: hwis01 systemd[]: Starting Restore LIO kernel target configuration...
Nov  :: hwis01 systemd[]: Started Restore LIO kernel target configuration.

5.允许tcp3206

配置防火墙或者nsg，允许tcp3206的访问。

这里就不展开了。

三、 在iscsi客户端配置iscsi

1.发现并login iscsi的disk

发现：

[root@hwic01 iscsi]# iscsiadm -m discovery -t sendtargets -p 10.1.1.5
10.1.1.5:, iqn.-.hw.is02:disk01.lun0

Login：

[root@hwic01 iscsi]# iscsiadm --mode node --targetname iqn.-.hw.is02:disk01.lun0 --portal 10.1.1.5 --login
Logging in to [iface: default, target: iqn.-.hw.is02:disk01.lun0, portal: 10.1.1.5,] (multiple)
Login to [iface: default, target: iqn.-.hw.is02:disk01.lun0, portal: 10.1.1.5,] successful.

此时在如下文件中，就有了相关iqn的信息：

[root@hwic01 /]# ls /var/lib/iscsi/nodes
iqn.-.hw.is01:disk01.lun0 iqn.-.hw.is02:disk01.lun0

此时在dev中已经存在sdc和sdd两个新增加的disk。

2.安装multipath软件

[root@hwic01 dev]# yum install device-mapper-multipath -y

复制配置文件：

cd /etc
cp /usr/share/doc/device-mapper-multipath-0.4./multipath.conf .

修改配置文件：

vim multipath.conf

blacklist {
devnode "^sda$"
devnode "^sdb$"
}
defaults {
find_multipaths yes
user_friendly_names yes
path_grouping_policy multibus
failback immediate
no_path_retry fail
}

启动服务：

[root@hwic01 etc]# systemctl enable multipathd
[root@hwic01 etc]# systemctl start multipathd
[root@hwic01 etc]# systemctl status multipathd
multipathd.service - Device-Mapper Multipath Device Controller
Loaded: loaded (/usr/lib/systemd/system/multipathd.service; enabled; vendor preset: enabled)
Active: active (running) since Fri -- :: UTC; 11s ago
Process:  ExecStart=/sbin/multipathd (code=exited, status=/SUCCESS)
Process:  ExecStartPre=/sbin/multipath -A (code=exited, status=/SUCCESS)
Process:  ExecStartPre=/sbin/modprobe dm-multipath (code=exited, status=/SUCCESS)
Main PID:  (multipathd)
CGroup: /system.slice/multipathd.service
└─ /sbin/multipathd
 
Nov  :: hwic01 systemd[]: Starting Device-Mapper Multipath Device Controller...
Nov  :: hwic01 systemd[]: PID file /run/multipathd/multipathd.pid not readable (yet?) after start.
Nov  :: hwic01 systemd[]: Started Device-Mapper Multipath Device Controller.
Nov  :: hwic01 multipathd[]: mpatha: load table [  multipath     service-time    :  : ]
Nov  :: hwic01 multipathd[]: mpatha: event checker started
Nov  :: hwic01 multipathd[]: path checkers start up

刷新multipath：

[root@hwic01 etc]# multipath -F

查看：

[root@hwic01 etc]# multipath -l
mpatha (36001405acadb3f79a2d44f48cafde627) dm- LIO-ORG ,disk01
size=.0G features='' hwhandler='' wp=rw
`-+- policy='service-time 0' prio= status=active
|- ::: sdc : active undef running
`- ::: sdd : active undef running

可以看到，两个iscsi的disk已经合并成一个disk了。

[root@hwic01 /]# cd /dev/mapper/
[root@hwic01 mapper]# ll
total
crw-------.  root root ,  Nov  : control
lrwxrwxrwx.  root root  Nov  : mpatha -> ../dm-

对dm-0进行分区：

[root@hwic01 dev]# fdisk /dev/mapper/mpatha
Welcome to fdisk (util-linux 2.23.).
Changes will remain in memory only, until you decide to write them.
Be careful before using the write command.
Device does not contain a recognized partition table
Building a new DOS disklabel with disk identifier 0x44f032cb.
Command (m for help): n
Partition type:
p primary ( primary,  extended,  free)
e extended
Select (default p): p
Partition number (-, default ):
First sector (-, default ):
Using default value
Last sector, +sectors or +size{K,M,G} (-, default ):
Using default value
Partition  of type Linux and of size  MiB is set
 
Command (m for help): p
 
Disk /dev/dm-:  MB,  bytes,  sectors
Units = sectors of  *  =  bytes
Sector size (logical/physical):  bytes /  bytes
I/O size (minimum/optimal):  bytes /  bytes
Disk label type: dos
Disk identifier: 0x44f032cb
 
Device Boot Start End Blocks Id System
mpatha1     Linux

格式化：

[root@hwic01 mapper]# mkfs.ext4 /dev/mapper/mpatha1

挂载、查看：

[root@hwic01 mapper]# mkdir /iscsi
[root@hwic01 mapper]# mount /dev/mapper/mpatha1 /iscsi
[root@hwic01 mapper]# df -h
Filesystem Size Used Avail Use% Mounted on
/dev/sda1 30G .2G 29G % /
devtmpfs 829M  829M % /dev
tmpfs 839M  839M % /dev/shm
tmpfs 839M 8.3M 831M % /run
tmpfs 839M  839M % /sys/fs/cgroup
/dev/sdb1 69G 53M 66G % /mnt/resource
tmpfs 168M  168M % /run/user/
/dev/mapper/mpatha1 985M 2.5M 915M % /iscsi

四、 检查iscsi disk的HA及其它属性

1. HA检查

将iscsi server1上的target服务停止掉：

[root@hwis01 target]# systemctl stop target
[root@hwis01 target]# systemctl status target
target.service - Restore LIO kernel target configuration
Loaded: loaded (/usr/lib/systemd/system/target.service; enabled; vendor preset: disabled)
Active: inactive (dead) since Fri -- :: UTC; 7s ago
Process:  ExecStop=/usr/bin/targetctl clear (code=exited, status=/SUCCESS)
Process:  ExecStart=/usr/bin/targetctl restore (code=exited, status=/SUCCESS)
Main PID:  (code=exited, status=/SUCCESS)
 
Nov  :: hwis01 systemd[]: Starting Restore LIO kernel target configuration...
Nov  :: hwis01 systemd[]: Started Restore LIO kernel target configuration.
Nov  :: hwis01 systemd[]: Stopping Restore LIO kernel target configuration...
Nov  :: hwis01 systemd[]: Stopped Restore LIO kernel target configuration.

在iscsi客户端查看：

[root@hwic01 iscsi]# multipath -l
mpatha (36001405acadb3f79a2d44f48cafde627) dm- LIO-ORG ,disk01
size=.0G features='' hwhandler='' wp=rw
`-+- policy='service-time 0' prio= status=active
|- ::: sdc : failed faulty running
`- ::: sdd : active undef running

可以看到一条路径已经出现故障。但磁盘工作仍然正常：

[root@hwic01 iscsi]# ll
total
-rw-r--r--.  root root  Nov  : a
drwx------.  root root  Nov  : lost+found
[root@hwic01 iscsi]# touch b
[root@hwic01 iscsi]# ll
total
-rw-r--r--.  root root  Nov  : a
-rw-r--r--.  root root  Nov  : b
drwx------.  root root  Nov  : lost+found

再恢复服务：

[root@hwis01 target]# systemctl start target
[root@hwis01 target]# systemctl status target
target.service - Restore LIO kernel target configuration
Loaded: loaded (/usr/lib/systemd/system/target.service; enabled; vendor preset: disabled)
Active: active (exited) since Fri -- :: UTC; 3s ago
Process:  ExecStop=/usr/bin/targetctl clear (code=exited, status=/SUCCESS)
Process:  ExecStart=/usr/bin/targetctl restore (code=exited, status=/SUCCESS)
Main PID:  (code=exited, status=/SUCCESS)
 
Nov  :: hwis01 systemd[]: Starting Restore LIO kernel target configuration...
Nov  :: hwis01 systemd[]: Started Restore LIO kernel target configuration.

客户端multipath仍然正常工作：

[root@hwic01 iscsi]# multipath -l
mpatha (36001405acadb3f79a2d44f48cafde627) dm- LIO-ORG ,disk01
size=.0G features='' hwhandler='' wp=rw
`-+- policy='service-time 0' prio= status=active
|- ::: sdc : active undef running
`- ::: sdd : active undef running

2. iops

根据前面的一片iops测试的文章中的方法，进行iops的测试：

[root@hwic01 ~]# ./iops.py /dev/dm-
/dev/dm-, 1.07 G, sectorsize=512B, #threads=, pattern=random:
 B blocks: 383.1 IO/s, 196.1 kB/s ( 1.6 Mbit/s)
 kB blocks: 548.5 IO/s, 561.6 kB/s ( 4.5 Mbit/s)
 kB blocks: 495.8 IO/s, 1.0 MB/s ( 8.1 Mbit/s)
 kB blocks: 414.1 IO/s, 1.7 MB/s ( 13.6 Mbit/s)
 kB blocks: 376.2 IO/s, 3.1 MB/s ( 24.7 Mbit/s)
 kB blocks: 357.5 IO/s, 5.9 MB/s ( 46.9 Mbit/s)
 kB blocks: 271.0 IO/s, 8.9 MB/s ( 71.0 Mbit/s)
 kB blocks: 223.0 IO/s, 14.6 MB/s (116.9 Mbit/s)
 kB blocks: 181.2 IO/s, 23.7 MB/s (190.0 Mbit/s)
 kB blocks: 137.7 IO/s, 36.1 MB/s (288.9 Mbit/s)
 kB blocks: 95.0 IO/s, 49.8 MB/s (398.6 Mbit/s)
 MB blocks: 55.4 IO/s, 58.1 MB/s (465.0 Mbit/s)
 MB blocks: 37.5 IO/s, 78.7 MB/s (629.9 Mbit/s)
 MB blocks: 24.8 IO/s, 103.8 MB/s (830.6 Mbit/s)
 MB blocks: 16.6 IO/s, 139.2 MB/s ( 1.1 Gbit/s)
 MB blocks: 11.2 IO/s, 188.7 MB/s ( 1.5 Gbit/s)
 MB blocks: 5.7 IO/s, 190.0 MB/s ( 1.5 Gbit/s)

可以看到，这个盘的iops大约在500左右，带宽在1.5Gbps左右。

总结：

通过Azure的File Service可以通过文件创建iscsi disk的方式把File Service中的文件发布成Disk，供iscsi客户端挂载。在多个iscsi Server提供ha服务的情况下，通过multipath软件可以实现HA的iscsi disk方案。

这个方案适合于一些集群需要共享磁盘、仲裁盘的情况。