Remember that you are not a System
Administrator. You should talk to your System Administrator to apply patches. In the case, if you
dont have a System Administrator do the following steps.
Steps to apply UNIX patches:
|
1 Download a most recent patch into
your secure server.
Go to http://www.sun.com
to get the most up to date patch applicable to your machine.
|
2 ftp your patch to the server that
needed to be upgraded.
Login to the server
you want to apply patches. Then check that patches have already applied and check the available
space.
$showrev p | grep patchnumber # to check that patch was applied or not.
$df k #write
down the place you want to replace the patch for example /u02
$mkdir /u02/patches # make directory to store the recommended patches
$ ftp the_secure_server
$ftp> bin
$ftp> get 8_Recommended.zip /u02/patches/8_Recommended.zip
Check the ftp downloaded file
size.
|
3 Unzip or untar the patch.
$cd /u02/patches
$unzip 8_Recommended.zip
--OR--
$tar
-xvf 8_Recommended.tar
Print and read
CLUSTER_README
$more /u02/patches/8_Recommended/CLUSTER_README
|
4 Check available disk spaces on
root, /var, /opt, and /usr.
$df k
check /, /usr, /var,
and /opt. Make sure that you have enough space on those file system.
|
4b Read the cluster patches
nstallation if you have cluster server (HA)
Skip this part if
you dont have cluster server or HA.
$Stop your clusters on both server.
|
5 Check disk layout for root disk.
$df k
Write down the file
system for root, /usr, /var, and /opt.
Or Print
$lp d your_printer_destination your_file_system for
ex: lp d comproom myfile
|
6 Stop
all databases or application running on the server.
|
7 Boot the server from
CDROM.
$su
# sign on as root
$init 0
ok boot cdrom s
|
8 Save all system information to a
tape.
User usfdump or any other
backup utility you have to dump the root, /var, /opt, and /usr file systems.
$mount /dev/dsk/c0t0d0s0 /a
$cd /a
Write a script to
dump all above file systems to tape.
$ufsdump 0f /dev/rmt/0n
/dev/dsk/c0t0d0s0
# for example as root
$ufsdump 0f /dev/rmt/0n
/dev/dsk/c0t0d0s2
# for example as /var
$ufsdump 0f /dev/rmt/0n
/dev/dsk/c0t0d0s3
# for example as /opt
$ufsdump 0f /dev/rmt/0
/dev/dsk/c0t0d0s4
# for example as /usr
use "vi
myufsdump" and write all ufsdump commands.
Save and close the
file.
Then execute the
script.
TERM=ansi;export TERM #
to user vi
$. ./myufsdump
Check the ufsdump to
be sure that your backup was successful.
$ufsrestore tvf /dev/rmt/0
|
9 Boot the server as a single user.
$init 0
$boot -s
|
10 Apply patches to the server.
Mount to the file system that
you save and unzip your patches. Then execute the install_cluster.
$mount /dev/dsm/c0t9d0s0 /u02
$cd /u02/patches/8_Recommended
$./install_cluster
|
11 Check the logs.
$more /var/sadm/install_data/
..log
|
12 Start
all databases or application software, and clusters on the servers, if there are no errors.
|
If you dont have enough space and
need to change or swap the file systems, you need to do the following.
Assuming you want to change /opt to /var
and /var to /opt.
Assuming also that /dev/dsk/c0t0d0s1
mounted as /var and /dev/dsk/c0t0d0s5 mounted as /opt.
$mt rewind
$mt fsf 1 Assuming this is the
place you have already dump /var to tape.
$newfs
/dev/rdsk/c0t0d0s5 clear /opt.
$mount /dev/dsk/c0t0d0s5 /a
$cd /a
$ufsrestore rvf
/dev/rmt/0n # restore /var and copy it into /opt location.
$ls al
--make sure it is
Repeat the same process for restoring
/opt in /var location.
|