> I'm curious, why can't you do a master/slave setup? It's just not all that useful for this particular application. Indexing new docs and merging segments - which as I understand is the main strength of having a write-only master - is a relatively small part of our app. What really is expensive perf-wise for us is the facet regeneration after each commit, and long-running complex queries. I'm not sure a master/slave setup would solve the issue with many open searchers - even with a long poll interval, won't the slave still potentially open many searchers if old searchers are still in use?
My current plan is to create a request handler that iterates through the mbeans in req.getCore().getInfoRegistry() and gets a count of how many distinct SolrIndexSearchers there are, then returns this count. Seems really hacky, but fairly straightforward. Would love to hear any ideas on a better way to do this. -Michael