My guess would be that the smaller limit causes large facet refinement requests to be sent out on the second phase. It's not clear what's happening after that though (i.e. why that causes things to crash)
-Yonik On Thu, Jun 2, 2016 at 8:47 AM, Markus Jelsma <markus.jel...@openindex.io> wrote: > Hello, > > I ran accros an awkward situation where is collect all ~7.000.000 distinct > values for a field via facetting. To keep things optimized and reduce memory > consumption i don't do setFacetLimit(-1) but a reasonable limit of 10.000 or > 100.000. > > To my surprise, Solr just stops or crashes. So, instead of decreasing the > limit, i increased the limit to a 1.000.000! And it works! The weird thing is > that with a limit of 100.000 or 200.000 and a heap of 3.5 GB, Solr stops. But > with a limit of 1.000.000 and a reduced heap of 2.5 GB, it just works fine. > > When it fails, it some times doesn't crash, but throw > > 396882 WARN (NIOServerCxn.Factory:0.0.0.0/0.0.0.0:9983) [ ] > o.a.z.s.NIOServerCnxn caught end of stream exception > EndOfStreamException: Unable to read additional data from client sessionid > 0x155111cc413000d, likely client has closed socket > at > org.apache.zookeeper.server.NIOServerCnxn.doIO(NIOServerCnxn.java:228) > at > org.apache.zookeeper.server.NIOServerCnxnFactory.run(NIOServerCnxnFactory.java:208) > at java.lang.Thread.run(Thread.java:745)s a > > This is on Solr 6.0.0 in cloud mode. 3 shards and 2 replica's on my local > machine. > > What is happening here? > > Many thanks, > Markus
