SLT Replication Issue Due To Timeout Reading Logging Tables
In our company we use SAP LT (Landscape Transformation) replication server to replicate tables from SAP ECC system to SAP HANA in real time. SLT creates trigger for each table that we want to replicate and whenever there is change it stores the changes in the logging table. Later on, these tables are used to update replicated tables in SAP HANA.
This system is really useful when you really need real time replication to SAP HANA. However, it also causes performance issue if proper monitoring is not being performed. Just recently, we faced an issue where we were trying to replicate a large table "VBFA" to SAP HANA in production environment. Every time table replicate status changes to "ERROR" after the initial load. It started to fail when ever SLT tries to read change/delta stored in logging table. This tables replication worked fine in DEV and QA environment.
After troubleshooting and taking help from SAP support we found that in the program generated by SLT (an ABAP program is generated every time we add a table for replication) contains SELECT statement which is used to read data from logging table did not have MANDT (Client) selected as part of the WHERE clause. MANDT is also part of the key fields in logging table. Therefore, every time it was causing timeout for SELECT statement (there seems to be issue with indexing or duplicate records). So as per the suggestion from SAP support we implemented the SAP note 1945151 .
Now onwards whenever we add new table for replication then the program module generated will contain MANDT in SELECT statement which also provides greater performance when reading delta. Table VBFA was already in system and there was already a program generated for that table (without MANDT in SELECT statement). So in order to fix that we used "Reset Replication Object Definition" which is located under Expert Functions in LTR cockpit (use transaction LTRC). This generated a new program (with MANDT in SELECT statement) for that table.
I hope this will help someone facing similar issue.