OK, turned out ZkStateReader.constructState() was only calling
ClusterState.getCollections()
for log.debug(). I removed that and the next bottleneck is talking
to ZkStateReader.fetchCollectionState.
"coreZkRegister-4-thread-14-processing-n:ftet1:8003_solr
x:t_1558_shard1_replica1 s:shard1 c:t_1558
Hi All,
I've had a first look at porting the patch I did for SOLR-7191 (SolrCloud
with thousands of collections) in Solr 4.10 to the Solr trunk (1708905).
Now I created 6,000 collections (3 nodes; 2 x replicas) and re-started the
3 nodes. What I noticed is that the cloud is starting but slowly. Al
