Dear All, Our SolrCloud cluster(4 nodes, 4 shards, Embedded Zookeeper) failed to start after VMs we started after weekend. We shut down 4VM in our private cloud for weekend and started SOLR in the same order as they were initialized - first zookeeper-hosting node and then 3 other nodes.
Unfortunately last node instead of being recognized as 4th shard was attached to shard 1 as a replica. This ended up with "no servers hosting shard" error - our cluster was corrupted and we have to clean it up. Our guess is that this mess happened due to reassigning of IP addresses to VM's after restarting. Do you have any guesses what could be a possible reason for this? -- Alex.
