Repartition Disk in Solaris using LU

Pre-Implementation Procedure:
==================================================================
1. Test & ensure console is accessible on server014

2. Test & ensure root password is correct

3. Check with Storage team & ensure latest OS full backup was taken successfully

4. take backup of
/etc/vfstab

5. take a snapshot of the following output:
metastat
metastat -p
metadb -i
df -k
Original Partition table:
Part Tag Flag Cylinders Size Blocks
0 root wm 825 - 1443 6.01GB (619/0/0) 12597888
1 swap wu 0 - 824 8.01GB (825/0/0) 16790400
2 backup wm 0 - 14086 136.71GB (14087/0/0) 286698624
3 unassigned wm 1444 - 1446 29.81MB (3/0/0) 61056
4 unassigned wm 1447 - 2065 6.01GB (619/0/0) 12597888
5 var wm 2066 - 2684 6.01GB (619/0/0) 12597888
6 unassigned wm 2685 - 2839 1.50GB (155/0/0) 3154560
7 unassigned wm 2840 - 14086 109.15GB (11247/0/0) 228898944

6. Ensure HIP is in test mode

7. Ensure live upgrade packages are installed:
pkginfo -l SUNWlur
pkginfo -l SUNWluu
** Skip #8 if live upgrade is installed **

8. If live upgrade not installed, install the packages from jumpstart server:
** depends on what Solaris version the system is running, for bresduxrac014, its running Solaris 9, so **

mount server4:/export/jump/Solaris_9/OS/Solaris_9_2005-09HW/Solaris_9/Product /mnt

pkgadd -d . SUNWadmap
pkgadd -d . SUNWlur
pkgadd -d . SUNWluu


** Following steps are for RAC servers only **
9. Shutdown patrolagent
/etc/init.d/patrolagent stop

10. Shutdown Rac cluster:
hastop -local

11. rename vcs startup script to prevent auto-start during reboot:
mv /etc/rc3.d/S99vcs /etc/rc3.d/s99vcs



Implementation Procedure:
===================================================================
1. Login as root via console

2. Break the root mirror (c1t1d0)
metadetach d1 d11
metadetach d2 d21
metadetach d3 d31
metadetach d4 d41
metadetach d5 d51
metadetach d6 d61
metaclear d11 d21 d31 d41 d51 d61
metadb -d c1t1d0s3
** test reboot after break mirror if necessary - shutdown -y -g0 -i6 **

3. Re-partition root mirror (c1t1d0) using the following size:
Part Tag Flag Cylinders Size Blocks
0 root wm 825 - 4122 32.01GB (3298/0/0) 67120896
1 swap wu 0 - 824 8.01GB (825/0/0) 16790400
2 backup wm 0 - 14086 136.71GB (14087/0/0) 286698624
3 unassigned wm 4123 - 4125 29.81MB (3/0/0) 61056
4 unassigned wm 4126 - 4744 6.01GB (619/0/0) 12597888
5 var wm 4745 - 6393 16.00GB (1649/0/0) 33560448
6 unassigned wm 6394 - 6806 4.01GB (413/0/0) 8405376
7 unassigned wm 6807 - 14086 70.65GB (7280/0/0) 148162560


4. Create boot environment on c1t0d0 disk:
lucreate -c abe -m /:/dev/dsk/c1t0d0s0:ufs \
-m -:/dev/dsk/c1t0d0s1:swap \
-m /var:/dev/dsk/c1t0d0s5:ufs \
-m /local:/dev/dsk/c1t0d0s7:ufs \
-m /var/crash:/dev/dsk/c1t0d0s4:ufs \
-m /opt/patrol:/dev/dsk/c1t0d0s6:ufs \
-n obe
** This may take a while for creating the Boot environment, please be patient and do not send any break key **


5. Verify that new LiveUpgrade boot environment creation is completed:
lustatus
Boot Environment Is Active Active Can Copy
Name Complete Now On Reboot Delete Status
-------------------------- -------- ------ --------- ------ ----------
abe yes yes yes no -
obe yes no no yes -

6. Take shapshot of the following:
lufslist abe
lufslist obe

** backout checkpoint #1 **

7. Activate obe to become next active boot environment & perform test boot against it:
luactivate obe ** take snapshot of the output of this command for backout purposes **
shutdown -y -g0 -i6 ** if reboot not successful, use the command taken from previous **
** snapshot to boot into the previous boot environment & backout **

8. After reboot, make sure obe is now active:
lustatus
Boot Environment Is Active Active Can Copy
Name Complete Now On Reboot Delete Status
-------------------------- -------- ------ --------- ------ ----------
abe yes no no yes -
obe yes yes yes no -

** root filesystem should have been grown from 6gb to 32gb **
df -k /

9. Take shapshot of the following:
lufslist abe
lufslist obe

10. Remove abe after it has became none active, and release the disk (c1t0d0) to become root mirror of obe:
ludelete abe

11. Make sure abe is no longer existed:
lustatus
Boot Environment Is Active Active Can Copy
Name Complete Now On Reboot Delete Status
-------------------------- -------- ------ --------- ------ ----------
obe yes yes yes no -

** backout checkpoint #2 **

12. Remove the old metadevices
metaclear d1 d2 d3 d4 d5 d6
metaclear d10 d20 d30 d40 d50
metadb -f -d c1t0d0s3

13. Reconfigure metadevices for the current boot device:
metadb -f -a -c 3 c1t1d0s3
metainit -f d11 1 1 c1t0d0s0
metainit -f d21 1 1 c1t0d0s5
metainit -f d31 1 1 c1t0d0s1
metainit -f d41 1 1 c1t0d0s6
metainit -f d51 1 1 c1t0d0s7
metainit -f d61 1 1 c1t0d0s4
metainit -f d1 -m d11
metainit -f d2 -m d21
metainit -f d3 -m d31
metainit -f d4 -m d41
metainit -f d5 -m d51
metainit -f d6 -m d61
metaroot d1

14. take backup & edit /etc/vfstab and change to metadevices:
cp -pr /etc/vfstab /etc/vfstab_bk
vi /etc/vfstab
** update the following entries **
/dev/md/dsk/d3 - - swap - no -
/dev/md/dsk/d2 /dev/md/rdsk/d2 /var ufs 1 no logging
/dev/md/dsk/d5 /dev/md/rdsk/d5 /local ufs 2 yes logging
/dev/md/dsk/d4 /dev/md/rdsk/d4 /opt/patrol ufs 2 yes logging
/dev/md/dsk/d6 /dev/md/rdsk/d6 /var/crash ufs 2 yes logging

15. Rename the following files:
mv /etc/lutab /etc/lutab_bk
rename any ICF.* found:
ls -l /etc/lu/ICF.*
if ICF.1 then
mv /etc/lu/ICF.1 /etc/lu/ICF.1_bk
if ICF.2 then
mv /etc/lu/ICF.2 /etc/lu/ICF.2_bk

16. Make sure no more boot environment existed:
lustatus ** should get no BEs found error **

17. Reboot server to ensure it can boot up using metadevices:
shutdown -y -g0 -i6

18. After server rebooted successfully, capture the following snapshots:
metadb -i
metastat
metastat -p
df -k -F ufs ** should only show metadevices **
lustatus ** should be none **
lufslist ** should be none **

19. Repartition root disk using the new partition size created on root mirror in step #3
prtvtoc /dev/rdsk/c1t1d0s2 | fmthard -s - /dev/rdsk/c1t0d0s2

prtvtoc /dev/rdsk/c1t0d0s2 (s) | fmthard -s - /dev/rdsk/c1t1d0s2 (d)

20. configure metadevices on the following disk slices (c1t0d0):
metadb -a -c 3 c1t1d0s3
metainit d10 1 1 c1t1d0s0
metainit d20 1 1 c1t1d0s5
metainit d30 1 1 c1t1d0s1
metainit d40 1 1 c1t1d0s6
metainit d50 1 1 c1t1d0s7
metainit d60 1 1 c1t1d0s4

21. Attach the metadevices as the submirror:
metattach d1 d10
metattach d2 d20
metattach d3 d30
metattach d4 d40
metattach d5 d50
metattach d6 d60

22. Monitor & make sure both metadevices are 100% sync'd :
metastat | grep "%"
** This may take a while for the metadevices to be in sync'd, do not reboot the server before the resync task fully completed **

23. After metadevices are 100% sync'd, turn vcs start-up script back on & perform final reboot to ensure server is back to normal state:
mv /etc/rc3.d/s99vcs /etc/rc3.d/S99vcs
shutdown -y -g0 -i6

** backout checkpoint #3 **


Post Implementation Procedure:
===================================================================
1. After reboot, perform verification to ensure all are running ok including VCS and all its resource groups:
uptime
df -k -F ufs
hastatus -sum



Back-out Plan:
===================================================================
1. At ** backout checkpoint #1 **
Repartition c1t1d0 back to the original size taken during Pre-Implementation Procedure# 5

2. At ** backout checkpoint #2 **
Reboot server back to its previous boot environment using the snapshot taken on Implementation Procedure# 7

3. At ** backout checkpoint #3 **
Repartition c1t0d0 & c1t1d0 disk back to the original size taken during Pre-Implementation Procedure# 5, and request Storage team for tape recovery

4. Reconfigure the metadevices as per the backup taken during Pre-Implementation Procedure# 5

5. Reattach root mirror

6. Execute Implementation Procedure #15 & 16 to clear off the boot environment.

** Back-out #7 is for RAC servers only **
7. Turn on VCS and reboot
mv /etc/rc3.d/s99vcs /etc/rc3.d/S99vcs

8. shutdown -y -g0 -i6