Lesson 4 of 5
Running and testing the synchronization
11 min
Listen to this lesson
Synchronization is where BPs are actually created. It runs in each ECC system, development first, then quality, then production, and in production it is a planned event with a business freeze on customer and vendor maintenance. Testing proves two things: that every record got a BP, and that each BP is correct.
The CVI load
| Section | Activity | Why |
|---|---|---|
| Pre-load | Deactivate interfaces (for example to a CRM system) | Avoid triggering outbound processing during the load |
| Pre-load | Block maintenance transactions | Stop users changing customers and vendors during the load |
| Real load | Load tool (MDS_LOAD_COCKPIT): customer to BP, vendor to BP | Create BPs and links; select by number, range or account group; run in background with parallel processes |
| Real load | Post-processing | Review and fix errors via the post-processing office (MDS_PPO2) |
| Real load | Completeness check | Confirm no customer, vendor or contact person is left unsynchronized |
| Post-load | Reactivate interfaces; release maintenance | Return the system to the business |
The CVI Load tab in sequence, as shown in SAP's training.
SAP's advice is to start with a small batch of five to ten customers or vendors, check the Monitor tab, then scale up. Errors appear with a red status; the "Call PPO" function shows the details, for example a tax number category that does not exist. Fix the source record and rerun for those numbers. After the load, CVI stays active in ECC: new or changed customers and vendors get or update their BPs automatically, and failures land in the PPO, so MDS_PPO2 must be checked regularly until conversion.
- Week 1ED1 loadSmall batch, then all; customizing defects fixed
- Week 3EQ1 loadFull volume rehearsal with runtime measured; test cycle on synchronized data
- Week 6EP1 loadWeekend with maintenance freeze; completeness check; sign-off
- Until conversionOperate with CVIDaily PPO check; upgrade preparation tab before SUM
- Red status in the Monitor tab?Yes → open the error via Call PPO · No → continue with the next batch
- Is the cause a data error in the record?Yes → owner corrects the customer or vendor; rerun for those numbers · No → next question
- Is the cause a missing mapping or customizing entry?Yes → fix customizing in development, transport, rerun · No → next question
- Unclear or repeating error?Yes → check the CVI notes and cookbook; open an incident if needed · No → document and close
Measure the runtime of the full load in EQ1 and use it to plan the production weekend, including time for post-processing and the completeness check. The number of parallel processes depends on the infrastructure; Sam tunes it in EQ1, not in production.
Testing the result
| Test | How | Expected result |
|---|---|---|
| Completeness | Completeness check; counts of customers, vendors and contact persons against BPs and CVI links (for example tables CVI_CUST_LINK and CVI_VEND_LINK) | Zero unsynchronized records; counts reconcile |
| Numbering | Sample by account group | BP numbers as decided (same or new range) |
| Roles | Sample per account group | Roles from the mapping, e.g. FLCU00 and FLCU01 for sold-to customers |
| Data correctness | Field comparison of name, addresses, tax numbers, bank details for a risk-based sample | Values equal to the source record |
| Category | Sample of records that should be persons | Category person where the derivation rule applies |
| Contact persons | Sample of customers with contacts | Contact person BPs with relationships to the organization |
| Ongoing sync | Create and change a customer and a vendor in ECC after the load | BP created or updated; nothing in the PPO |
| Processes | Sales order, delivery and billing for Great Lakes Fleet Services; PO, invoice and payment for Lakeside Components | Processes run unchanged in ECC with CVI active |
Contoso's synchronization test set.
Choose samples by risk, not at random only: every account group, every country, records that failed and were corrected, merged customer-vendor pairs if you merge, one-time accounts, and the top customers and suppliers by revenue and spend. For bank details and tax numbers of high-value partners, compare every record, not a sample.
Worked example: reconciling Contoso’s counts
| Source (EP1) | Count | Target | Count after load | Result |
|---|---|---|---|---|
| Customers in KNA1 | 4,850 | Links in CVI_CUST_LINK | 4,850 | Reconciled |
| Vendors in LFA1 | 1,960 | Links in CVI_VEND_LINK | 1,960 | Reconciled |
| Customer contact persons in KNVK | about 2,300 | Contact person links (CVI_CUST_CT_LINK) | Same as source | Reconciled after two reruns in EQ1 |
Contoso’s completeness evidence (fictional figures). Count with SE16N or a query, and store the outputs with the date and system.
In EQ1 the first count did not match: CVI_CUST_LINK held 4,838 rows against 4,850 customers. The 12 gaps were all one-time customers whose account group had no role mapping. The fix was customizing, so it went through ChaRM change control in ED1, was transported, and the 12 were rerun. That is the pattern to expect: gaps cluster by account group or country, which is why samples are drawn per account group.
What would you do?
After the EQ1 load, the completeness check shows 18 vendors not synchronized. The load must move to EP1 next week. What should happen?
- Load in sequence: pre-load, load tool, post-processing, completeness, post-load.
- Start small, check the Monitor and PPO, then scale with parallel processes.
- Test completeness, numbering, roles, data, category, contacts, ongoing sync and processes.
- Keep checking MDS_PPO2 until conversion; overall cockpit status must be green.
Check your understanding · not graded
Lesson check
1. A key test after CVI synchronization is:
2. Which transaction shows CVI errors in the post-processing office?
3. What does SAP recommend before loading all records?
4. Why block maintenance transactions during the load?
5. After the EQ1 load, KNA1 holds 4,850 customers but CVI_CUST_LINK holds 4,838 rows. What does that tell the test lead?