Tek-Tips is the largest IT community on the Internet today!

Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!

  • Congratulations derfloh on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!

Is boot -s needed for replacing failed mirror disk? 1

Status
Not open for further replies.

gallows

Technical User
Joined
Jun 12, 2004
Messages
223
Location
US
From reading the posts it looks like "tis the season for bad disks".

My root mirror failed and I need to replace the disk and remirror the boot disk. According to Sun, they recommend performing this process in Single User Mode. Is this really necessary? I hate to bring a production server down for a couple of hours if I don't have to. My thinking is that the Single User mode is a CYA for Sun.

BTW, OS is Solaris 9, 280R and using Sun's Disk Manager.

Thanks,
Gallows
 
I have replaced the root as well as the mirrored root many times with the O/S running.

This is an example from a VX80 , the E280 is similar due to the internal fiber drives.

By the way did you use the following setting in your /etc/system file: You can boot with less than 50% of the DB replicas.

set md:mirrored_root_flag=1


***************HOW TO REPLACE AN INTERNAL FIBER DISK ON A Vx80***************************************************


Determine the bad disk by looking at the messages and metastat ouput

On this example c1t0d0 (BOOT DISK) on elsaa02 is used




1. run metadb -i and look at the replicas, see how many and the status
2. metadb -d c1t0d0s6 (this will delete the replicas on slice 6
3. luxadm remove_device /dev/rdsk/c1t0d0s2
4. devfsadm -C
5. look at cfgadm -al
6. When prompted, remove the bad disk
7. ls /dev/rdsk/c1t0d* (do you see any links)
8. Install the new disk
9. luxadm insert_device c1t0d0s2
10. format (do you the new disk)
11. ls /dev/rdsk/c1t0d* (do you see the links)
12. cfgadm -al
13. prtvtoc /dev/rdsk/c1t1d0s2 | fmthard -s - /dev/rdsk/c1t0d0s2
14. metadb -a -f -c2 c1t0d0s6
15. metadb -i
16. installboot/usr/platform/`uname -i`/lib/fs/ufs/bootblk/dev/rdsk/c1t0d0s0
17. metareplace -e d0 c1t0d0s0
metareplace -e d1 c1t0d0s1
metareplace -e d7 c1t0d0s7
18. metastat |grep -i progress
21. when resyncing is done you must make the new disk bootable (ALSO AT THE FIRST OPPORTUNITY TEST THE BOOTS)


EXAMPLE ONLY: INFO IS FROM THE OLD "BAD DISK"
22. Determine the device path to the boot devices for both the primary and mirror:

ls -l # ls -l /dev/dsk/c1t0d0s0 /dev/dsk/c1t1d0s0
lrwxrwxrwx 1 root root 70 Jan 24 2003 /dev/dsk/c1t0d0s0 -> ../..
/devices/pci@9,600000/SUNW,qlc@2/fp@0,0/ssd@w21000004cfcb0cfa,0:a
lrwxrwxrwx 1 root root 70 Jun 15 12:06 /dev/dsk/c1t1d0s0 -> ../..
/devices/pci@9,600000/SUNW,qlc@2/fp@0,0/ssd@w2100002037429cc5,0:a


Use the device paths to define the sds-root and sds-mirror device aliases (NOTE USE THE LABEL "disk" INSTEAD OF "ssd" IN THE DEVICE
PATH)

eeprom "nvramc=devalias sds-root /pci@9,600000/SUNW,qlc@2/fp@0,0/disk@w21000004cfcb0cfa,0:a devalias sds-mirror /pci@9,600000/SUNW,qlc@2/fp@0,0/disk@w21000004cfc57f81,0:a

eeprom "use-nvramrc?=true"

eeprom (look at boot device)

TEST THE BOOT AT THE EARLIEST POSSIBLE TIME)
 
Outstanding reply Tuka!
Thanks very much. I thought that I had done this before but wasn't 100% sure.

Gallows
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor

Back
Top