Hi all, I have noticed the following occur with some consistency: When I execute a long running query (that spans 15 or more seconds), the Solr node that is servicing the request starts to perform a full copy from the shard leader. My current configuration has only one shard with 3 replicas. Note that there are no updates happening on any of the Solr nodes in the cluster; hence there really shouldn't be any changes to the underlying index, nor is there any need to synchronize the index with other replicas.
I'm just trying to understand what kind of events can typically trigger syncing the index with other replicas...specifically resulting in a full copy, when no updates have been made. Could this have to do with some timeout value settings for Zookeeper...where the Solr server is unable to respond to a heartbeat request to report it's state? Any pointers would be greatly appreciated. Thanks! -Niran
