Project procedure (Cluster 3.0)

jackeyzhang

1. Backup OS
  
     Shutdown Cluster and backup to tape , before do this , run explorer and keep results .
     # scshutdown -y -g0
     Ok. boot -x
     #ufsdump 0uf /dev/rmt/0hn /
     #ufsdump 0uf /dev/rmt/0hn /oracle
     #ufsdump 0uf /dev/rmt/0hn /export/home
     
2. Restore OS from 18GB to  73GB disk .

    format and label 73GB disk , according to poson doc for the new disk partition
   
    ０ ／（根目录） 20GB
1 ／Swap 区 20GB
2 系统保留
3 vm private
4 vm public
5 /oracle 8GB
6 /export/home 22GB
7 /globaldevices 100M
合计  约70G

    I recommend the globaldevices partition to 512mb (this is for SC3.1 need)
   
   newfs 73GB disk and restore OS
   #mount /dev/dsk/c1t0d0s0 /mnt
   #cd /mnt
   #ufsdump 0uf - / | ufsrestore rvf -
   
   Restore the oracle and opt , home partition to 73GB disk .
   Newfs the global device partition
   #newfs /dev/rdsk/c1t0d0s7
  
3. mount the 73GB root and change some file
    #mount /dev/dsk/c1t0d0s0 /mnt
    #cd /mnt/etc
    #touch /mnt/etc/vx/reconfig.d/state.d/install-db
    Remove the flag that tells Volume Manager that the root disk is encapsulated:
      
    # rm /mnt/etc/vx/reconfig.d/state.d/root-done
    #cp /mnt/etc/system /mnt/etc/system.orig
    #vi /mnt/etc/system
删除或者采用*注释其中两行
* rootdev:/pseudo/vxio@0:0
* set vxio:vol_rootdev_is_volume=1
#cp /mnt/etc/vfstab /mnt/etc/vfstab.orig
#vi /mnt/etc/vfstab
  编辑vfstab文件，将vfstab文件还原为没有封装前的状态，修改文件中所有以以下一行
原有：
/dev/vx/dsk/swapvol     -       -       swap    -       no      -
/dev/vx/dsk/rootdg/rootvol /dev/vx/rdsk/rootdg/rootvol / ufs 1 no logging
/dev/vx/dsk/rootdisk_17vol      /dev/vx/rdsk/rootdisk_17vol     /global/.devices/node@1 ufs     2       no      global
修改为：
/dev/dsk/c1t0d0s1     -       -       swap    -       no      -
/dev/dsk/c1t0d0s0   /dev/rdsk/c1t0d0s0   / ufs 1 no logging
/dev/dsk/c1t0d0s7      /dev/rdsk/c1t0d0s7     /global/.devices/node@1 ufs     2       no      global
node2参考node1
/dev/dsk/c1t0d0s7      /dev/rdsk/c1t0d0s7     /global/.devices/node@2 ufs     2       no      global
     
  4. Install bootblk and boot the 73GB disk
   # umount /mnt
   # cd /usr/platform/sun4u/lib/fs/ufs;installboot ./bootblk /dev/rdsk/c1t0d0s0
   #init 0
   Ok. boot 73GBdisk -x
   
  5. Restore the c0t0d0 (18GB) partition to original
    When Volume Manager encapsulates a disk, it makes a record of the old vtoc of the disk.
This file is stored for each disk in /etc/vx/reconfig.d/disk.d/c#t#d#.  This file is stored in a Volume Manager
specific format, so it can't be used as an argument to fmthard unless it is modified.  The 'vxedvtoc' command is
similar to fmthard but knows how to read this vtoc file and write that vtoc to a disk.  The command takes the form:
      vxedvtoc -f  
      
Assuming that the boot disk is c0t0d0 we would now run the command
      /etc/vx/bin/vxedvtoc -f /etc/vx/reconfig.d/disk.d/c0t0d0/vtoc /dev/rdsk/c0t0d0s2
      
# THE ORIGINAL PARTITIONING IS AS FOLLOWS :
#SLICE     TAG  FLAGS    START     SIZE
0         0x0  0x200        0        0
1         0x0  0x200        0        0
2         0x5  0x201        0  8794112
3         0x0  0x200        0        0
4         0x0  0x200        0        0
5         0x0  0x200        0        0
6         0xe  0x201        0  8794112
7         0xf  0x201  8790016     4096
# THE NEW PARTITIONING WILL BE AS FOLLOWS :
#SLICE     TAG  FLAGS    START     SIZE
0         0x0  0x200        0  2048000
1         0x0  0x200  2048000  2048000
2         0x5  0x201        0  8794112
3         0x0  0x201  4096000  2048000
4         0x0  0x201  6144000  2048000
5         0x0  0x200        0        0
6         0x0  0x200        0        0
7         0x0  0x200        0        0
DO YOU WANT TO WRITE THIS TO THE DISK ? [Y/N] :y
WRITING THE NEW VTOC TO THE DISK
   About the c0t1d0s0 (old 18GB mirror disk)  , just remove the slice 3 and 4 ( it keep the old rootdg info)
   
6. Vxinstall and encapsulate 73GB disk

    Encapsulate root via vxinstall
When answering the prompts for the name of the
rootdisk for each node - enter a different name for
each node.
e.g.: rootnode0 rootnode1 ...
Note the unique minor numbers in the cluster e.g.:
Reminor the device in rootdg to be unique on each node. Set each node to
     have a base minor number of 100*. E.G. node1=100, node2=200
     vxdg reminor 00
Check via: sccheck command.
   #boot -r
  调整系统minor值
当系统minor值在两台主机上冲突时候，需要调整其中一台主机的minor值。
# vxdg reminor rootdg 100(node1)
#vxdg reminor rootdg 200 (node2)
Reminor the device in rootdg to be unique on each node. Set each node to
     have a base minor number of 100*. E.G. node1=100, node2=200
     vxdg reminor 00
修改完成后reboot 出现冲突的主机
7.  待两台主机都启动cluster后上运行
# devfsadm
# devfsadm －C
更新DID信息
# scgdevs
清除多余的DID信息
# scdidadm  －C
# scgdevs
检查DID信息
# scdidadm －L   
其中共享硬盘会后重复的两行，本地硬盘只有一行。
将机器reboot一次检查是否由DID报错信息。并用scstat 检查锁盘状态是否online
8. Mirror root disk

   vxdiskadm init another 73GB disk name rootmir rootmir

9. checked the vxdmp disable or not

   #vxdisk list
   If display as "fabricx" , vxdmp is disabled
   If display as "t3_0" , vxdmp is enable . must run vxdiskadm to disable it .
   
   
  10. Get the latest explorer output to me for check  

本文来自ChinaUnix博客，如果查看原文请点：http://blog.chinaunix.net/u/14743/showart_105138.html