AIX upgrade fails

Hello Team,

Am trying to upgrade the AIX 6.1 TL 7 to TL 8. My rootvg is mirrored so i have unmirrored and taken out the secondary disk and trying to install the update via smitty alt_clone.

The filesets has been installed and finally when making the updated disk as bootable bosboot is failing. Please help me ...

 
 -------------------------------------------------------------------------------
 bos.aso                     6.1.8.16        USR         APPLY       SUCCESS
 bos.aso                     6.1.8.16        ROOT        APPLY       SUCCESS
 bos.aso                     6.1.8.16        USR         COMMIT      SUCCESS
 bos.aso                     6.1.8.16        ROOT        COMMIT      SUCCESS
 install_all_updates: Log file is /var/adm/ras/install_all_updates.log
 install_all_updates: Result = SUCCESS
 Modifying ODM on cloned disk.
 Building boot image on cloned disk.
  
 0301-150 bosboot: Invalid or no boot device specified!
 usage:  bosboot {-a | -v} [-d device] [-p proto] [-k kernel] [-l lvdev]
                 [-b file] [-M primary|standby|both] [-D|-I] [-LTq]
         Where:
         -a              Create boot image and write to device or file.
         -v              Verify, but do not build boot image.
         -d device       Device for which to create the boot image.
         -p proto        Use given proto file for RAM disk file system.
         -k kernel       Use given kernel file for boot image.
         -l lvdev        Target boot logical volume for boot image.
  
  
  
  
  
 fcapdev[/home/root]#ln -f /dev/rhdisk0 /dev/ipldevice
 ln: 0653-419 /dev/rhdisk0 and /dev/ipldevice are identical.
 fcapdev[/home/root]#bosbood -ad /dev/hdisk0
 ksh: bosbood:  not found.
 fcapdev[/home/root]#bosboot -ad /dev/hdisk0
  
 0301-168 bosboot: The current boot logical volume, /dev/hd5,
         does not exist on /dev/hdisk0.
 fcapdev[/home/root]#ipl_varyon -i
 [S 10616942 7471268 11/16/13-00:46:57:783 ipl_varyon.c 1265] ipl_varyon -i
  
  
 PVNAME          BOOT DEVICE     PVID                    VOLUME GROUP ID
 hdisk0          NO              0008761ae1775e910000000000000000        0008761a0000d400
 hdisk1          YES             0008761a49a99d340000000000000000        0008761a0000d400
 [E 10616942 0:097 ipl_varyon.c 1405] ipl_varyon: exited with rc=0
 fcapdev[/home/root]#lslv -m hd5
 hd5:N/A
 LP    PP1  PV1               PP2  PV2               PP3  PV3
 0001  0001 hdisk1
  

R/G,
Gowtham

What is the output of lspv ?

What steps you did to unmirror rootvg?

Did you applied the bundle?

Can you paste oslevel -s output?

When you un-mirrored, do you expect to be able to reboot from the removed disk? This is not the case. If you did something like:-

unmirrorvg rootvg hdisk1
reducevg hdisk1
rmdev -dl hdisk1

... and then pulled it, then hdisk1 is empty if the reducevg completes.

It seems counter intuitive, but assuming that the two disks were completely mirrored and bootable, you would have been safer to shutdown, pull the disk, boot. That way, if you need to, you could have forced the boot via SMS to start from the disk you pulled out.

If hd5 is not on hdisk0, this would suggest that it was only on hdisk1. The unmirror will not move LVs that are not mirrored. This would mean that you could not take hdisk1 out the VG of course, so that would suggest that something was wrong at that point.

Before we panic too much. Are you happy that you have recent mksysb media to recover with?

Robin

I suggest you do an "alt-disk-install" - which would be a fresh install of the new version of the OS - or an "alt-disk-migration", which means:

  • breaking the mirror of a mirrored rootvg
  • update the OS on the broken-off mirror
  • restart the system (from the newly updated disk)
  • resynchronize by remirroring using the disk with the old OS version

alt-disk-installation is quite similar, just instead of an update a fresh install is done.

Fortunately all this is done by the procedure itself, you do not do (manually) any of these steps. I just wrote it so you can picture what goes on during the procedure.

Put the disk back in, establish the mirror again and then start over doing an alt-disk-migration. For a procedure for performing an alternate disk migration read the IBM infocenter page.

I hope this helps.

bakunin

1 Like

I would like to make a point here, for Alt_Disk_Install, he actually needs a empty and free disk.

You don't have to place the disk back and mirror.

Before we do Alt_Disk_Install, we actually try to break the mirror and then use that empty disk to clone the OS. It is a backup copy of the current OS

After you done you can either upgrade this or alt_disk image.

But if you want to take a backup of existing image and upgrade the binaries at the same time you use alt_disk_clone.

Your current disk will server as backup disk (it will stay at current version even after reboot) and other empty disk you use for alt_disk_clone operation will be upgraded to the new desired OS version. In both cases you just have to change the bootlist to jump different version (ofcourse after a reboot).

1 Like

You are right and i should have clarified what i mean: my suggestion was to do an alternate disk migration, which needs the mirrored disk in place. It will break up the mirror as described in the process, but the procedure does that itself, you don't have to provide the broke-up mirror yourself.

I hope this helps.

bakunin

Sorry again, but if I am recalling it right for NIM ADM (which we perform from NIM server).
We need a free target disk. Correct me if I am wrong.