See also from the migration section of Infocenter at
Preparing for the migration from a System i model that does not support
IBM i 7.1
Right below the link for "Restoring previous release user data to a new
system" you'll see
The preferred method to restore previous release user data on your new,
target system is to use the migration method.
The migration method asks you to first install the new, current release
onto your old, source system. After which, you save your old system and
then you perform a full system recovery onto the new, target system. Only
use these instructions if it is not possible to perform the preferred
migration process referenced in the Data migrations topic.
Notice there's a lot of manual recreation using the "Restoring previous
release user data to a new system" steps instead of just upgrading your
source system to 7.1 first. For example this is one of them from the
Add job scheduler entries with the Add Job Schedule Entry (ADDJOBSCDE)
command using the information you printed from your source system. Use the
Work with Job Schedule Entries (WRKJOBSCDE) command and select the print
Now, we have a few hundred job schedule entries. Granted some might say
why not just restore the *JOBSCD object? Question this, why didn't IBM
just tell you that? Did they reformat that at some release that would
have been addressed if you had done a normal migration (by updating your
source system first)? Let me give you an example. Let's say you go from
2.1 of BPCS to 8.3. You decide to skip all the migration steps that BPCS
had for data conversions in between. Do you think you could just restore
the 2.1 item master (IIM) file and have it work?
I'm not saying it's impossible to upgrade this way; evidently numerous
people have. I am just saying that you may not end up saving yourself any
work and there's been a number of people on this list that's taken over a
year to clean up a mess caused by doing it this way.
The OP came from Mount Olive Pickles. I would suspect there might be a
slow time of the year in which they could do this over two weekends. OS
upgrade one weekend and the hardware upgrade on another. There's a
company called Red Gold (a tomato processor) which attends our local user
group and I am just going by some of the stuff I've heard from them.
If 24x7x365 is really a necessity then I'd argue they really should have a
HA environment and should switch over to their backup machine and stagger
the upgrades that way.