Migrating from VSE to Linux

We are now using Adabas/Natural on IBM/VSE.

As support is ending in 2024 we have to migrate to another platform - preferably Linux.

Are there anyone out there who has tried that - or anyone who can point to a roadmap describing the steps necessary to achieve that goal.

I would be thankful for any hints and advices.

There have been many successful ports from IBM mainframe to LUW. No idea how many were VSE.

Search these forums for “rehost”. Then search SAG-L, the listserv hosted by the University of Arkansas.

Porting Adabas files is easy (unless your users won’t accept ASCII collating sequence or you mix packed values into your alpha fields). It just takes a bit of work. Adabas provides all the utilities you need. 99% of your Natural code will move cleanly, but there are a few syntax incompatibilities, a few coding techniques that are problematic, and ASCII conversion.

Getting help can be difficult because those with the most experience in this area (independent contractors and SAG professional services) hesitate to give this advice for free.

The areas that cause the most difficulty are usually not the Adabas/Natural pieces, but such things as other code (COBOL or Assembler), job control, remote job entry and printing management. There are some third-party tools that can assist with this migration such as CICS emulators to support COBOL/CICS code and tools to migrate JCS/JCL to scripts.

I did a migration like this a few years ago. It was a very interesting experience. But, I don’t see any difficulty in this. The VSE helps a lot in this process. If you need anything, I’m at your disposal.

Thank you all.
It sounds promising that most of the migration is out-of-the-box.
I dont think that EBCDIC to ASCII will be the big challenge.
Most of out middelwaresystems are ASCII based anyway.

The challenges will be CICS and JCL handling and jobscheduling, but I think that we will overcome that.
I would like to mark all 3 as Accepted Answer/Solution - but I’m not sure it’s doable.

I don’t think that EBCDIC to ASCII will be the big challenge.

Unless your programmers routinely coded alphanumeric ranges as

' ' THRU '9'
'A' THRU '9'

You will need to change them to

' ' THRU 'z'
'0' THRU 'z'
'0' THRU H'FF'

Or they may have coded hex values, such as H’05’ for TAB.

1 Like

If at all possible, replace any IF statements/WHERE clauses using ranges with IS or MASK instead as it is more portable and clearer as to what the purpose is.

For the same reason as Ralph mentions, with database access statements, be aware that "FROM ‘A’ " or "FROM ‘0’ " may give a different result on Linux than it did on the mainframe if the data being read contains both alphas and numerics.

1 Like

I fully agree. And I have full confidence in the programmers abilities