v880上做RAID5与恢复的方法

本文档以SUN Fire V880 Server（6*73GB Dsik）为例讲解在Solarsi8下使用命令行的方式安装及维护RAID1和RAID5。本文档内容都经本人实践验证过，但对于不同机型仅供参考，如需转载请注明出处，谢谢支持！ 

Disk分布情况： 
0:c1t0d0 
1:c1t1d0 
2:c1t2d0 
3:c1t3d0 
4:c1t4d0 
5:c1t5d0 

注：0,1做镜像；2，3，4，5做RAID5 

系统盘(c1t0d0)分区及镜像盘(c1t1d0)情况： 
c1t0d0s0(d10)      d0      c1t1d0s0(d20)        /               1024M 
c1t0d0s1(d11)      d1      c1t1d0s1(d21)        swap            8192M 
c1t0d0s2                  　　c1t1d0s2            overlap          69999M 
c1t0d0s3(d13)      d3      c1t1d0s3(d23)        /usr             4096M 
c1t0d0s4(d14)      d4      c1t1d0s4(d24)        /opt             10240M 
c1t0d0s5(d15)      d5      c1t1d0s5(d25)        /var             2048M 
c1t0d0s6(d16)      d6      c1t1d0s6(d26)        /home           free 
c1t0d0s7                  　　c1t1d0s7            /metaDB         30M 

系统安装完成后，使用Solarsi8 software 2 of 2安装DiskSuite_4.2.1软件，安装路径：/cdrom/cdrom0/Solaris_8/EA/installer & 选择默认安装就可以。 
用root用户登陆,运行以下命令： 
#prtvtoc /dev/rdsk/c1t0d0s2|fmthard –s – /dev/rdsk/c1t1d0s2 
fmthard:New volume table of contents now in place 
上面命令将第二块硬盘(c1t1d0)的文件分区表调整为和引导盘一致。 

#umount /metaDB 
#rm –r /metaDB 
#vi /etc/vfstab 
将下面这行注释掉或删除,如下： 
#/dev/dsk/c1t0d0s7 /dev/rdsk/c1t0d0s7 /metaDB ufs 1 yes –   

一、RAID1 
对各个分区逐一作镜像： 
①先生成replicas,这是DiskSuite内部用的。 
#metadb –a –f –c 3 c1t0d0s7 c1t1d0s7 
#metadb 
②Creating a mirror from swap 
#metainit –f d11 1 1 c1t0d0s1 
#metainit d21 1 1 c1t1t0s1 
#metainit d1 –m d11 
#vi /etc/vfstab 
/dev/dsk/c1t0d0s1 - - swap – no – 
should be changed to: 
/dev/md/dsk/d1 - - swap – no - 
#reboot 
#metattach d1 d21 
③Creating a mirror from /usr 
#metainit –f d13 1 1 c1t0d0s3 
#metainit d23 1 1 c1t1t0s3 
#metainit d3 –m d13 
#vi /etc/vfstab 
/dev/dsk/c1t0d0s3 /dev/rdsk/c1t0d0s3 /usr ufs 1 yes – 
should be changed to: 
/dev/md/dsk/d3 /dev/md/rdsk/d3 /usr ufs 1 yes – 
#reboot 
#metattach d3 d23 
④Creating a mirror from /opt 
#metainit –f d14 1 1 c1t0d0s4 
#metainit d24 1 1 c1t1t0s4 
#metainit d4 –m d14 
#vi /etc/vfstab 
/dev/dsk/c1t0d0s4 /dev/rdsk/c1t0d0s4 /opt ufs 1 yes – 
should be changed to: 
/dev/md/dsk/d4 /dev/md/rdsk/d4 /opt ufs 1 yes – 
#reboot 
#metattach d4 d24 
⑤Creating a mirror from /var 
#metainit –f d15 1 1 c1t0d0s5 
#metainit d25 1 1 c1t1t0s5 
#metainit d5 –m d15 
#vi /etc/vfstab 
/dev/dsk/c1t0d0s5 /dev/rdsk/c1t0d0s5 /var ufs 1 yes – 
should be changed to: 
/dev/md/dsk/d5 /dev/md/rdsk/d5 /var ufs 1 yes – 
#reboot 
#metattach d5 d25 
⑥Creating a mirror from /home 
#metainit –f d16 1 1 c1t0d0s6 
#metainit d26 1 1 c1t1t0s6 
#metainit d6 –m d16 
#vi /etc/vfstab 
/dev/dsk/c1t0d0s6 /dev/rdsk/c1t0d0s6 /home ufs 1 yes – 
should be changed to: 
/dev/md/dsk/d6 /dev/md/rdsk/d6 /home ufs 1 yes – 
#reboot 
#metattach d6 d26 
⑦Creating a mirror from / 
#metainit –f d10 1 1 c1t0d0s0 
#metainit d20 1 1 c1t1t0s0 
#metainit d0 –m d10 
#metaroot d0 
#lockfs –fa 
#reboot 
#metattach d0 d20 
      
#metastat  (检查镜像进度) 


镜像完成后，还需作如下操作： 
修改EEPROM 
ok devalias（察看启动设备） 
ok nvalias rootdisk /pci@8,600000/SUNW,qlc@4/fp@0,0/disk@0,0 
ok nvalias mirrdisk /pci@8,600000/SUNW,qlc@4/fp@0,0/disk@1,0 

ok setenv boot-device rootdisk mirrdisk 

eeprom命令： 
boot-device=rootdisk mirrdisk 
boot-device=rootdisk mirrdisk 
use-nvramrc?=true 
nvramrc=devalias mirrdisk /pci@8,600000/SUNW,qlc@4/fp@0,0/disk@1,0 
devalias rootdisk /pci@8,600000/SUNW,qlc@4/fp@0,0/disk@0,0 

#ls -l c1t0d0s0 
lrwxrwxrwx   1 root     root          70  1月  2 10:45 c1t0d0s0 ->; ../../devices/pci@8,600000/SUNW,qlc@4/fp@0,0/ssd@w21000000870e80e7,0:a 
对应：/pci@8,600000/SUNW,qlc@4/fp@0,0/disk@0,0 
root@voiplnjvr4 # ls -l c1t1d0s0 
lrwxrwxrwx   1 root     root          70  1月  2 10:45 c1t1d0s0 ->; ../../devices/pci@8,600000/SUNW,qlc@4/fp@0,0/ssd@w21000000870fb287,0:a 
对应：/pci@8,600000/SUNW,qlc@4/fp@0,0/disk@1,0 

测试： 
ok boot rootdisk 系统启动正常。 
ok boot mirrordisk 系统启动正常。 

假如c1t0d0盘坏掉，则停机更换硬盘后需做恢复工作，如下； 
ok boot mirrordisk -s 
#metadb -d c1t0d0s7 
#prtvtoc /dev/rdsk/c1t1d0s2|fmthard –s – /dev/rdsk/c1t0d0s2 
#metadb –a –f –c 3 c1t0d0s7 
#halt 
ok boot mirrordisk 
#metareplace -e d0 c1t0d0s0  
d0: device c1t0d0s0 is enabled  
#metareplace -e d1 c1t0d0s1  
d1: device c1t0d0s1 is enabled  
#metareplace -e d3 c1t0d0s3  
d3: device c1t0d0s3 is enabled  
#metareplace -e d4 c1t0d0s4  
d4: device c1t0d0s4 is enabled  
#metareplace -e d5 c1t0d0s5  
d5: device c1t0d0s5 is enabled  
#metareplace -e d6 c1t0d0s6  
d6: device c1t0d0s6 is enabled 

#metastat  (检查镜像恢复进度) 

二、RAID5 
#metainit d55 –r c1t2d0s2 c1t3d0s2 c1t4d0s2 c1t5d0s2 
#metastat d55 (检查RAID5的进度) 
作完后reboot 
#newfs /dev/md/rdsk/d55 
#mkdir /raid5 
#vi /etc/vfstab 
添加一行： 
/dev/md/dsk/d55 /dev/md/rdsk/d55 /raid5 ufs 2 yes – 
#reboot 


RAID5坏掉一个盘的恢复方法： 
例c1t4d0盘坏掉，更换坏硬盘后作如下工作： 
ok boot -r 
#metareplace –e d55 c1t4d0s2 
#metastat d55 

卸去RAID5的方法： 
#umount /raid5 
#metaclear d55 
#vi /etc/vfstab 
注释此行： 
#/dev/md/dsk/d55 /dev/md/rdsk/d55 /raid5 ufs 2 yes –