Today I noticed this on my dev server running Solr 5.2.1: https://www.dropbox.com/s/bt81sv35acb7q2n/searcher-from-old-corename.png?dl=0
The name of this core is spark0live, but before the last index rebuild, it was named spark0build. My full rebuild process indexes data into build cores, then when the rebuild is done, it swaps the build cores with the live cores. This core is holding on to the searcher object from before the core swap. The last rebuild I did was days ago, so it was getting pretty old. A quick test on a 5.3.1 download (not a full test with my config/schema) did not reveal any problems. When I did some manual testing with core swaps on the 5.2.1 dev server, I could not reproduce the problem. On the index with the problem, I am using one third-party plugin with a schema component and an update processor, and an update processor that I wrote myself. Is it possible that these might be causing this to happen if objects are not being closed properly? I've since restarted the server, so I will need to initiate a full rebuild to see if the problem comes back. Thanks, Shawn
