Monday, 22 August 2011

LPAR migration step by step

You can minimize few of the administration tasks on the steps you have

I hope you will have the NIM setup. If yes, try to use the very latest
mksysb image.

# Are you migrating your LPAR to the same POWER series or to the different
power series?

# (Like, P5 to P5 / P6 to P6 / P5 to P6 or P7 / P6 to P7 etc.)

Configure NIM resources.

Ensure you have the client defined on the NIM master.

Ensure you have mksysb resource created.

Note :- Based on the above point highlighted with # mark you can decide
whether you would need lpp source or not. For the drivers for new power

If it's same power series, you would not required lpp source. If not, lpp
source required for the SAME oslevel.

Ensure you have spot created

Check you have any requirement to put first boot script and do it
accordingly (optional).

1. Backup - Take the mksysb backup on NIM and TSM incremental backup.

2. DB's and application will be down.

3. Take one more incremental (So, it will cover the file changes happened
during the shutdown process and it won't take much time like the time it's
taken on the step1).

4. Shutdown the source server.

5. Inform to SAN team to map the disk's to target server, (give the WWPN's
which you have taken when you tested the LPAR with the temp ip) Note:- Do
not unmap the disk's from the source AS OF NOW. I will cover the reason at
the end.

6. Deallocate all the resources assigned to this client earlier.

7. Define the latest copy of mksysb image to the mksysb resource created for
the client.

8. Allocate the resources mksysb, spot, lppsource (if you're migrating to
different P series hardware), fb script.

nim -o allocate -a spot=spot_VR00_TLxxSPxx \
-a lpp_source=lpp_VR00_TLxxSPxx \
-a fb_script=firstboot \
-a mksysb=mksysb_<servername> <servername>

9. Perform lsnim -l <servername> and check all the resources are assigned

10. Initiate the bos installation for this server.

11. Check the /etc/bootptab file the entry added properly.

12. Check with the SAN team that have mapped the disk's to the target
server. If yes. step 13.

13. Activate the target server in SMS mode.

14. Perform IPL setting and test the connectivity with the NIM server
performing the ping test.

15. Boot the machine with network adapter.

16. On the installation menu, you will see all the data disk's along with
the rootvg disk.

The purpose that you have tested your target LPAR with the temp IP is to
take the PVID of the rootvg disk and the WWPN's of the target lpar - I am

17. Verify the PVID of the rootvg disk which you have taken earlier and
select the rootvg disk on the installation menu.

18. On the same menu you will have a option for import volume group. Select
yes for that, so when you machine comes up all the filesystem's will be
mounted and you can start verify the count's of the mount point's.

No need to manually import the VG.

19. Continue with the installation.

20. Once the server came up's check paging, aio, mount point's, etc.

21. If all seems good, inform to the DB and Application guys to start and
test it.

22. Change the boot mode of source server to Open firmware (Rename the LPAR
name for identification purpose saying Do not boot stuffs like that).

23. Wait for 2 to 3 days and unmap the disk's from the source server.

24. If the step 21 is not good or anything went wrong. Then proceed from
step 25.

25. Shutdown the target server and inform to the san team to unmap the data
disk's which is assigned to target server.

26. Boot the source server. It will become fastest roll back (continuation
of step 5).

Sorry for the long explanation ;-). You can take the points which are applicable for you.

1 comment:

Twitter Bird Gadget