The algorithm is only sensitive to the shard ID, you should be able to freely move the data to another node.
BTW, perhaps the easiest way to do this would be to set up a replica for the shards you care about on the new hardware (assuming connectivity) and let Solr do the synchronization for you. Once the new replica is up and running, you can kill off the node on the old machine and the new one will just take over. Best, Erick On Mon, Mar 17, 2014 at 7:08 AM, ku3ia <dem...@gmail.com> wrote: > Hi all! > A have a solrcloud (v. 4.6.0) cluster of 5 shards. Four of them I need to > move to another server (configs and indexes). So as I understand I need to > clear zookeeper data and after restart it will be updated. My question is, > for example, in a future I need to update a specific document, is a guaranty > that the SolrCloud internal distributing algorithm will work properly and > document will be routed to concrete shard and will be updated? Does the > algorithm depends of actual data location or not? > > Thanks. > > > > > -- > View this message in context: > http://lucene.472066.n3.nabble.com/Solrcloud-DistribIndexing-question-tp4124802.html > Sent from the Solr - User mailing list archive at Nabble.com.
