No CDC on clustered tables using SAP Application DB endpoint.

[Anonymous]

Hello,

We have created two tasks for testing. Both are logstream task. They connect to a SAP endpoint. One is using SAP Application and is working "fine", it reads the data and capture the new incoming changes. The bad thing about it is that as we are treating with clustered tables, it only apply changes one by one. So we are testing SAP Application DB.

The SAP Application DB endpoint on the other hand, reads the data from the source, so we are able to make a full load, but it does not detect any incoming changes (and there are incoming changes, as the other task, the SAP Application, is showing them).

What could be wrong here? We need the incoming changes from the SAP clustered tables.

Anything you need let me know.

[sureshkumar]

Hello Team,

While SAP Application we do some limitation

"Changes are applied to cluster tables is in transactional mode only. If you need the Batch Optimized Apply functionality, consider using the SAP Application (DB) endpoint instead"

https://help.qlik.com/en-US/replicate/May2022/Content/Global_Common/Content/SharedReplicateHDD/SAP-A... 

Regarding SAP Application DB, we need Diagnostic Package to understand the issue

or you can open support case to analyze further

https://community.qlik.com/t5/Support/ct-p/qlikSupport 

Regards,

Suresh

[Anonymous]

Hello, 

Here you have the diagnostic package.

[Steve_Nguyen]

1. from the DP, i see that backend Database is Oracle, why not just use Oracle endpoint instead of SAP App DB ?

2. from the log i do see that it is reading the source : if you search in the log for :

Current Redo log sequence is  , you will see that the redo log is changing that mean Replicate is reading the log. 

3. lower source_capture ,, trace and then enable following to trace

sorter ,, trace

target_apply ,, trace

then see any target_apply activity.

[Steve_Nguyen]

also, i see many of the following :

[ASSERTION       ]W:  Incorrect PART_NO 1 in the Cluster 'KOCLU' object part 0, possible changes are ignored 

[Anonymous]

Hello,

1. We have an Oracle connection endpoint stablished, and we have two endpoints, a SAP Application DB and SAP Application, both using the same backend endpoint (Oracle).

2. The task is reading the source, but I understad that once it reads the source, it filters and only get the changes for the tables inside the task. So the task is reading the logs, but is not getting any changes for the clustered tables. While in another task we have a SAP Application endpoint reading from the same database and it gets all the changes for those clustered tables. But we need to use SAP Application DB to use batch apply mode instead of transactional apply (apply changes one by one).

Do you have any idea why we are not able to read changes for the cluster tables?

[shashi_holla · Accepted solution]

The following warning has been fixed in new versions of Replicate 2021.11 Service pack 07 or 2022.05 SP 03:

Incorrect PART_NO 1 in the Cluster 'KOCLU' object part 0, possible changes are ignored 

Please upgrade to those versions and try the behavior.

Thank you,