IX5000恢复域的过程

IX5000恢复域的过程 

当IX5000域中所有存储控制模块出现了不可恢复的严重错误并导致域无法进入正常状态，此时唯一的办法就是更换存储控制模块。

第一步：中断所有业务；

第二步：将需要更换的SCM撤下同时安装并连接好新的单个SCM（如果有多个SCM需要更换，请先连接一个SCM恢复域之后再将其他SCM配置好网络地址后加入该域中）。

第三步：启动SCM，出现以下字符，在系统询问是否需要配置网络时选择Yes，配置好SCM的网络地址后，当系统询问是否需要配置域时选择No，进入StorControl>提示符下进入debug shell后输入scmadm stop命令停止slate进程。

StorControl (TM) 2.2.002 [Build 9997]

Copyright (C) 2002-2005 Huawei-3com, Inc. All Rights Reserved.

scm login: admin

Password: ********

       #----------------------------------此处省略中间的提示字符-------------------------------------#

Do you wish to use the Realm's Network Configuration Wizard? (Yes|No|Quit) [No]:Yes

       #----------------------------------此处省略中间的提示字符-------------------------------------#

Huawei-3com Command Line Interface

You have logged in as 'admin'

User role is 'admin'

StorControl> debug shell

root@scm:~:Debug> scmadm stop

Stopping slate.... on Mon Mar 20 03:32:21 GMT 2006

Waiting for slate to shutdown . Done.

第四步：运行/opt/intransa/diag/recover_scm.sh脚本开始进行域信息恢复。在出现Do you want to continue?的提示信息之后，关闭域内所有DE，并仅开启1个DE，此时磁盘数为16“

root@scm:~:Debug> diag/recover_scm.sh

This Realm recovery utility should only be run when

replacing the last Storage Controller Module of a Realm.

It must be run on a new Storage Controller Module with no

other Storage Controller Modules existing within the Realm.

Please contact Intransa Technical Support before continuing.

Make sure that StorControl server is NOT running.

Make sure you have powercycled all your Disk Enclosures

connected to this Storage Controller before continuing further.

Do you want to continue? (Yes|No) [No]: yes

Removing zombie pid file.....

DEBUG: Stopping udhcp and ntpd if running.

DEBUG: Starting udhcp (Range 10.0.3.2 10.0.3.250)...

udhcp server (v0.9.8-pre-intransa8) started

Enter the number of disks the realm has: 16

#-----------------------------此处省略提示信息N行-------------------------------#

mount /devfs/idevice/0 /opt/intransa/smfs -t ext3 -orw,sync,data=ordered

smfs management volume mounted (-orw,sync,data=ordered)...

IOE version of the current StorageController  = 2.2.002

BuildID of the Current StorageController      = bROSH3CR1_060317-000

IOE version of the previous StorageCOntroller= 2.2.002

BuildID of the previous StorageController     = bROSH3CR1_060317-000

IOEversion and BuildId in StorageController match that in the MgmtVolume.

**********************************************************************

*                                                                     *

*                                                                     *

*           PLEASE NOTE DOWN THE REALM NAME GIVEN BELOW      *

*               For configuring the system, use the realm name: realm202        *

*                                                                      *

*             Please powercycle all your Disk Enclosures. Failure to do so may    *

*           affect your realm availability upon reboot.                         *

*                                                                      *

*                                                                      *

************************************************************************

root@scm:~:Debug> #------powerclycle all the DEs---------#

root@scm:~:Debug> #------reboot the SC--------#

root@scm:~:Debug> reboot

此时将所有DE均下电后上电，重启SCM。

第五步：SCM重启完成后，开始域配置状态下，记住配置域名和原来的域相同。完成域配置后输入新的密码，这样原来域的信息就全部回来了：

Do you wish to configure a Realm on this node? (Yes|No|Quit) [No]: yes

Waiting for proper controller state:  Done.

The following sections will ask for a Realm name, a DNS domain and a

Realm Management IP address.  It is strongly suggested that Realm name

map to the Realm Management IP address in your DNS.  That is, if the

realm name is myrealm, the DNS domain is ***.*** and the Realm

Management IP Address is 10.20.30.40, we suggest that the host

myrealm.***.*** have the IP address 10.20.30.40 in your DNS.

Enter the Realm Name (ex. myrealm) [Not Configured]: realm202

第六步：将其余SCM连接好后配置网络地址加入已恢复的域。