Full reindex is using language configuration of default instance for all existing instances

• Type: Bug
• Status: Closed
• Resolution: Fixed
• Affects Version/s: 6.2.10 EE GA1, 6.2.X EE, 7.0.0 CE GA1, 7.0.1 CE GA2
• Fix Version/s: 6.2.X EE, 7.0.0 DXP FP1, 7.0.2 CE GA3, 7.0.0 DXP SP1, 7.0.3 CE GA4, Master
• Component/s: Search Infrastructure, Search Infrastructure > Framework
• Labels:
  • QA-NF
  • liferay-digital-enterprise-70-sp1
  • liferay-fixpack-de-1-7010
  • liferay-fixpack-portal-111-6210
• Branch Version/s:
  7.0.x, 6.2.x
• Backported to Branch:
  Committed
• Story Points: 1.5
• Fix Priority: 3
• Git Pull Request: https://github.com/brianchandotcom/liferay-portal/pull/40981

Full reindex is using language configuration of default instances for all existing instances.

As there are some indexed fields that are localized, sometimes we are storing unnecessary data and other times we are skipping necessary languages.

For example:

1. default instance: en_US, en_GB, es_ES
2. second instance: es_ES
3. third instance: de_DE

In all cases we will index the fields only for "en_US, en_GB, es_ES" so:

• second instance will have unnecessary "en_US, en_GB" languages stored
• third instance won't have "de_DE" indexed (and the other languages will be unnecessary indexed)

Steps to reproduce

1. Configure two instances:
   • default one with several languages (for example: en_US, ca_ES, es_ES)
   • second one with only one language (for example: es_ES)
2. language configuration of each instance is modified login into the instance and going to: Control Panel => Configuration => Instance Settings => Miscellaneous
3. Create a DDM structure in both instances and add some webcontents
4. Check the indexes of both instances with Luke/Elasticsearch: each instance will have JournalArticle objects with fields only in its configured locales
5. Execute a full reindex from Control Panel => Configuration => Server Administration => Resources
6. Check again the indexes of both instances with Luke/Elasticsearch:
   • Wrong behavior: default instance is fine but second instance JournalArticle objects has fields with the locales of first instance
   • Expected behavior: second instance JournalArticle objects only stores "es_ES" fields