Out of sync due to corruption in Shareplex_rowidmap before 5.1.7 (27631)

Table that contains chaining row go out of sync. please check event_log on the source and you may see error like "Internal error: 10653 - can't fix corruption [sp_ocap/2678]", also check capture log (ocap.log) for the similar error like "07-01-22 15:55:41.962260 2678 1 Can't find nrid (AAAGlnACQAAASslAAA) error, transop listing", which all indicate shareplex has problem to find next rowid for chaining information in Shareplex_rowidmap.

The out-of-sync is caused by corruption of the SHAREPLEX_ROWIDMAP. Capture processes uncommitted ROLLBACK operations as NULLs (TRP_NULL). The OPP module does this, but due to not reinitializing internal variables, the TRP_NULL remains and causes Capture to not update the ROWIDMAP for the following non-NULL operation. This missing update causes following activity on the rowid in question to generate incorrect DML, thus leading to out-of-syncs on the target.

The out-of-sync is caused by corruption of the SHAREPLEX_ROWIDMAP.

This is known issue existed prior 5.1.6. Please run compare without repair if the table is not that huge, then apply sql script generated from compare to sync up data. Also please upgrade Shareplex to 5.1.7.