When importing an object via xml import, without replacing existing objects, the Lucene index is being re-built using the modification date of the oldest object that's skipped during the import. This causes the Lucene index to take some time as it performs a refresh of all objects with a modification date far back.
Steps to reproduce
After importing the object the JCP writes the following message to the log file:
U00045253 Search index update: Because of a long running transaction all objects with a modification date of '<time_stamp>' UTC or newer are indexed again.
Release : 12.3.3 and 12.3.4
Component : AUTOMATION ENGINE
The issue has been confirmed to be a bug and will be fixed with a future service pack for Automation Engine V12.3 .