dsmiley commented on PR #4237: URL: https://github.com/apache/solr/pull/4237#issuecomment-4135270719
Thinking of the bigger picture of Solr's internal APIs, caching or not, I'm reminded of a year ago working closer with ClusterState. ClusterState has a bit of javadocs claiming that it is immutable. But it actually isn't -- it can fetch newer state. It's live nodes can change later. Some parts of Solr implicitly expect ClusterState to be an immutable snapshot of the world -- replica placement plugins and I think some general cluster processing. I'd rather see a simplification of ClusterState to _actually_ be immutable, and instead the ClusterStateProvider be the primary API to get / cache state.json & livenodes & aliases & cluster properties etc. Aligned with this would be moving some of CloudSolrClient's caching into an improved ClusterStateProvider base class. I wonder what you think @HoustonPutman ? -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
