AnsweredAssumed Answered

SOLR clustering and Index tracking

Question asked by udarora on Sep 14, 2015
I am trying to do a SOLR clustering in my local environment with 2 instances. I had successfully clustered it and it is working fine.

My current architecture is like I have 2 alfresco instances running on different ports i.e. 8080 and 9090. Both the instances are pointing to a common content store and a common DB running on postgres on port 5432. Both the instances are having separate SOLR servers running at ports 8443 and 9443 respectively. But instead of having local copies of indexes, I am pointing both the SOLR servers to a common index base at a different location. Things are working fine until now.

But now I want only one SOLR server to behave as an indexer and the other one as only a searcher. To achieve this, I disabled the index tracking on the 9443 server by setting enable.alfresco.tracking=false

It did worked for me and the indexing was done by only by one server and not by both of them. Though searching was working on both the instances as expected.

But here is the issue, now if I shutdown the application server which was acting as an indexer, the searching for the already indexed content on the other server also stops. I can understand that it should not index new content as the indexing is turned off, but it should be able to search the content that is already indexed.

Please let me know if I am missing anything.