ATG-Endeca Site outage: Front End application goes down when baseline_update.bat|sh or promote_content.bat|sh script is triggered (Endeca v3.1.0, Endeca v3.1.1 and Endeca 3.1.2)

Problem Statement: 
ATG-Endeca Site goes down when baseline_update.bat|sh or promote_content.bat|sh script is triggered. This happens on applications which uses Endeca v3.1.0, Endeca v3.1.1 and Endeca 3.1.2 

Solution:
User may not see any error message when baseline_update.bat|sh or promote_content.bat|sh script is triggered but ERROR message will be logged in IFCR log file with failed rule which is created in Experience Manager with a specific location.

Ex: *ERROR*  POST /ifcr/system/endeca/mdexPublisher HTTP/1.1] com.endeca.ifcr.publish.impl.PublishService Unable to publish config documents com.endeca.ifcr.configService.ConfigServiceException: Config service error: Validation failed with 1 error [Unknown DVAL_ID 100972 specified for MERCH_RULE_TRIGGER in rule with id 21. Ignoring this MERCH_RULE element.]
Quick fix will be Correct all invalid rule(s) by modifying the problem rule(s) and set a valid location for each using Experience Manager and re-promote the content using promote_content.bat|sh script

Permanent fix: 
Apply below patch which is strongly recommended by ORACLE.
16761103: PUBLISH ACTION FOR EXPERIENCE MANAGER FAILS ON INVALID DVAL IDS
 OR
upgrading to MDEX 6.4.1.2 or later which contains these fixes 

This patch alters the MDEX behavior so that if an Experience Manager rule is ignored due to a missing dimension value used as the trigger, the MDEX publish action will still proceed with the valid rules and succeeds instead of failing.   After applying the patch, you will need to re-promote the content to the engines so that the MDEX's will end up with a good set of rules/content.

2 comments: