Full-text Search in Cluster

[mrinmoy_khamrui]

I am evaluating Alfresco and as I was going through the documentation I had a doubt how full-text search works within a cluster in the following scenarios.

1. There are two nodes in a cluster - node-A and node-B. Both nodes share a shared content store and shared database. A new content got created by node-A and both metadata and full-text index got created and stored in node-A. Because of ehcache synchronization, the Lucene index for metadata is node-B as well. Now the question is does node-B creates the full-text index as well for the same content? If not any full text search request in node-B will fail eventually.

2. There are two nodes in a cluster each having their own primary content store and replicates the contents to a central secondary content store. In such scenario, as there is no way for the content which got created by node-A being accessible by node-B, it seems the full text search hitting node-B will not be able to give proper results.

I am little confused over these and any help will be highly appreciated.

Thanks
Mrinmoy Khamrui

[openpj]

1. Yes, the second instance will create indexes as well.Because all the Lucene indexes are indipendent for these instances: each instance has its own indexes.
The content store is shared and it is unique.

If not any full text search request in node-B will fail eventually.

No, Alfresco verify if that content exist, if this content searched doesn't exist, Alfresco search the same content in the cache to retrieve it from the other instance.
If this new content is found, then it will imported with its own indexes as well.

2. Indexing work in the same way as the first answer above.

Hope this helps.