请问光纤盘做系统盘恢复的问题!

cjp

e3500服务器,solaris8操作系统,由于系统盘出现问题(备有各个系统分区的ufsdump文件),我换了块同样的新盘,按原来的分区,并且restore了各个分区,/ 分区做了installboot,但是从硬盘不能启动,据说还要做其它的操作,比如把光纤盘的WWN号写入prom等,请问哪位做过,给予指点,不胜感谢!

cjp

告诉我到哪查资料也行, 很着急,最好给个连接,先谢谢!

Philmoon

ok boot disk的时候有没有说是invalid wwn什么的？如果有，重新设置一下disk，使它指向你现在的盘。就是你说的“写入PROM”。

cqwlyh

printenv看看boot-device这项后面的内容，如果必要，可以使用setenv boot-device指向你要引导的盘。

melonm

在OK下，做
OK>;setenv autoboot? false
OK>;reset-all
...
OK>;probe-scsi-all
检查boot-device，指向新认出的光纤盘。最后，把autoboot设为true。
OK>;setenv autoboot? true

scsi

英文的，慢慢看。
1) The boot disk fails and is replaced by another disk with a different
    WWN.

2) Boot the system into single user mode from the cdrom or network.

3) Label the replacement disk to match the slices from the failed disk.

4) Create filesystems on all the slices to be restored.

5) Mount the root filesystem and restore the data from backups.

6) Install the boot block onto the recovered root slice using the
   "installboot" command.  Refer to the installboot manpage for the syntax.

7) With root mounted under "/a" run the following commands to re-build
    the devices tree :

   /usr/sbin/drvconfig -r /a/devices -p /a/etc/path_to_inst

   cd /devices
   /usr/bin/find . -print | cpio -pduVm /a/devices

   /usr/sbin/disks -r /a
   /usr/sbin/devlinks -r /a

NOTE: It is currently necessary to use both "drvconfig" and "find | cpio"
due to bugid 4161768, drvconfig does not work properly with socal disks.

Configure the "boot-device" parameter in the EEPROM using the
   "luxadm set_boot_dev /dev/dsk/{root slice entry}" command.

9) Restore the other filesystems on that disk, or comment out the
    entries for them from /a/etc/vfstab. At the very least, you must have all the
    Solaris filesystems (root, /var, /usr, /opt, etc.) recovered.

10) Reboot the system from the recovered disk.

cjp

谢谢scsi,我按照你提供的文档成功替换了系统盘.