How to Migrate a Productive CRM System to the SAP HANA™ Platform

You might have picked up the news: SAP’s own CRM system now runs productively on the SAP HANA platform. In this blog, you will learn how SAP’s IT organization managed to migrate one of the world’s largest CRM systems to the SAP HANA platform in only 2.5 months. 

So how was this change managed? And what are the lessons learned? 

From a technical point of view, migration to the SAP HANA platform involves two steps: 

• “Upgrade” Installation of Enhancement Pack (EHP) EHP2 for SAP CRM 7.0 on HANA
• OS/DB migration to the latest SAP HANA DB release and revision

Both of these steps are well-known SAP practice and both have to be performed on one cutover weekend. However, some special preparations are required.

Before making any significant system change, it is important to do your housekeeping thoroughly, such as re-enforcing your archiving system and deleting unnecessary data (for example, change documents). This will keep your system lean and reduces the migration downtime.

  

Based on test upgrades and migration rehearsals, you can figure out how to deal with the largest tables during the migration. The largest tables can be split during the export, which will significantly reduce the migration runtime. Testing is always very important, namely functional tests of the upgraded system, backup/restore tests, and failover tests of the SAP HANA high availability solution (depending on the SAP HANA hardware vendor). For the SAP HANA migration, you need to focus on load and performance testing. This allows you to optimize the system for SAP HANA before going live. Some tables  perform better in column or row store, and some expensive statements need to be optimized.

  

Finally, you should consider performing at least two rehearsals of the entire procedure before production cutover.

Image may be NSFW.
Clik here to view.
The ramp-down procedure is a standard process: unscheduled long running jobs, block-out the end users, clean up queues, isolate the system, and finally shut down the system. Cutting the high availability cluster at this stage has some significant advantage: One side is used for the actual upgrade and migration while the other side remains original. This original system can be used as a quick failback in case something goes wrong. After successful migration, this original system can be used as a reference system (for audits, for example).

  

The next step is the “upgrade” installation of the EHP2 for SAP CRM 7.0 on SAP HANA (technical CRM 7.12 based on SAP NetWeaver 7.40). Since this is only a minor EHP upgrade, the EHP2 installation only took 3 hours.

  

The actual migration consists of a DB export followed by a SAP HANA 1.0 SP5 import (in the meanwhile export and import can run in parallel). Large tables are split during the export and cluster tables are de-clustered since SAP HANA does not support cluster tables. The export of 3 TB took us 7 hours, and the import to SAP HANA 12 hours including the declustering. During the migration runtime, the application servers can be upgraded for SAP HANA.

   

After successfully migrating, it is essential to carry out some checks to verify that all tables and rows have been migrated properly and the dictionary is consistent (package checker, table checker). It also makes sense to perform additional content checks for the most important data (such as your sales pipeline).  

The ramp-up requires minimal effort as there is hardly any post-processing. Before opening the system, a spot check allows you to check whether everything in the migrated system is correct. End users should be ramped up with the assistance of a support team.