Expand my Community achievements bar.

SOLVED

Clustered env - Search indexes

Avatar

Level 4

There are different indexing options provided by AEM. I am currently on AEM 6.1

  1. Lucene index - Stores index data under repository (crx-quickstart\repository\index). Index config resides on AEM (/oak:index)
  2. Solr index - This would store index data on remote Solr Instance, configurations are on AEM side (/oak:index)
  3. External solr without indexing config on AEM - This would have data indexed manually through code to remote solr instance and then search hits can directly be made to solr instance. (Here we don't have to worry about clustering on the AEM side)

First 2 options search hits happen through AEM through querybuilder. Question is

In case of clustered environment does each publish instance needs to have index configuration? And each AEM would have indexed data stored? Can multiple publish instances share the index data?

Thanks
Sandeep

1 Accepted Solution

Avatar

Correct answer by
Level 9

Hi Sandeep,

     Firstly let me correct your understanding  & then update the solution.  Index will be stored under /oak:index in the repository itself along with definition.  Just that It is hidden.  The one you see under crx-quickstart\repository\index is copy from repository to make read access fast. 

In clustering(mongodb,...)  it will be same index defnition and index shared across all instance just that they have own copy in local.   I would recommend enable on read  so that each instance have own copy to make read fast.  If need you can turn off at http://host:port/system/console/configMgr/org.apache.jackrabbit.oak.plugins.index.lucene.LuceneIndex...  

In case of cold standby each instance will have own copy of index defnition and data.   

Thanks,

View solution in original post

1 Reply

Avatar

Correct answer by
Level 9

Hi Sandeep,

     Firstly let me correct your understanding  & then update the solution.  Index will be stored under /oak:index in the repository itself along with definition.  Just that It is hidden.  The one you see under crx-quickstart\repository\index is copy from repository to make read access fast. 

In clustering(mongodb,...)  it will be same index defnition and index shared across all instance just that they have own copy in local.   I would recommend enable on read  so that each instance have own copy to make read fast.  If need you can turn off at http://host:port/system/console/configMgr/org.apache.jackrabbit.oak.plugins.index.lucene.LuceneIndex...  

In case of cold standby each instance will have own copy of index defnition and data.   

Thanks,