GuestbookSign our guestbook ContactGet in touch with the authors ArchiveAll unixwerk articles since 2003
August 17, 2012

Tip: Alternate Disk Installation in Phases



  1. Introduction
  2. Phase 1
  3. Phase 2
  4. Phase 3
  1. Related Information


1. Introduction

Alternate Disk Installation is a nice method to update a system's TL or SP with minimal downtime. But there are flaws - for example if your installation contains interim fixes a TL upgrade will fail. And, how can you check for completeness of the installed service pack. Further, how can you avoid another reboot if interim fixes have to be applied to the system?

Running the three phases of the alternate disk installation separately does the trick!


2. Phase 1

The first phase of the alternate disk installation procedure is the cloning part. The target disks (hdisk2  and  hdisk3  in the below example) must not belong to a volume group.

 # alt_disk_install -C -P 1 "hdisk2 hdisk3"

After the cloning procedure has completed the alt_rootvg ist still opened and the filesystems are mounted. So we can work on the alternate rootvg before we continue with the installation - we will use this option to remove all interim fixes from the installation. This can be done with the chroot command:

 # INUCLIENTS=1 chroot /alt_inst /usr/sbin/emgr -l

 === ===== ========== ================= ========== ======================================
 1    S    IV13820s   04/10/12 15:59:33            Ifix for IV13820 @ AIX 61 TL0 SP05.   

 # INUCLIENTS=1 chroot /alt_inst /usr/sbin/emgr -r -L IV13820s

Please note the environment variable INUCLIENTS in front of the chroot command. Setting this variable to any non-empty value prevents installp and other commands from restarting services, changing device attributes, and other things. If you forget to set the variable services on your active rootvg might be stopped from installation scripts. You probably don't want that.


3. Phase 2

The second phase of the alternate disk installation will apply fixes to the alternate rootvg if requested. You can run Phase 2 multiple times in a row what puts you in the position to first update the installation to a new TL, then apply a SP, and finally install interim fixes. The alternate rootvg remains online after phase 2.

 # alt_disk_install -C -P 2 -b "update_all" -I "acNgXY" -l /install/tl06 "hdisk2 hdisk3"

In the above command /install/tl06 is the location where you stored the TL's update filesets.

Once the TL update has finished you can go on applying the last service pack to the alternate rootvg:

 # alt_disk_install -C -P 2 -b "update_all" -I "acNgXY" -l /install/sp08 "hdisk2 hdisk3"

In the above command /install/sp08 is the location where you stored the service pack update filesets.

At this point you can already check the OS Level of the alternate rootvg:

 # chroot /alt_inst /usr/bin/oslevel -s

Interim fixes cannot be installed with the alt_disk_install command. But you can install the fixes the same way we removed the fixes in phase 1. Please note that the fix packages have to be located in the alternate rootvg - otherwise they can't be accessed in the chroot environment. This is different to the TL/SP upgrade with alt_disk_install where the update filesets are outside the alternate rootvg! If the fix packages are not already in the alternate rootvg we can just copy them over:

 # cp /install/*.epkg.Z /alt_inst/tmp

and install them with the same trick as we used in phase 1:

 # INUCLIENTS=1 chroot /alt_inst /usr/sbin/emgr -e /tmp/IV21379.120613.epkg.Z
 # INUCLIENTS=1 chroot /alt_inst /usr/sbin/emgr -e /tmp/IV24827s08.120806.epkg.Z


4. Phase 3

The third phase finally will close the alternate rootvg and change the bootlist to the alternate rootvg. If you don't want the bootlist to be already changed you can use the  '-B'  flag:

 # alt_disk_install -C -P 3 [-B] "hdisk2 hdisk3"


A Related Information