You should run replicated ZK: http://zookeeper.apache.org/doc/trunk/zookeeperStarted.html#sc_RunningReplicatedZooKeeper Give Solr the list of all ZK's and you're good to go
-- Jan Høydahl, search solution architect Cominvent AS - www.cominvent.com Solr Training - www.solrtraining.com 7. feb. 2013 kl. 21:55 skrev mizayah <miza...@gmail.com>: > Hey, > > I just wonder is there any way to tell solr node that my zookeper instance > is down and new one is on another ip? > > What I want to achive is to have one zookeeper instance for claster of solr > nodes. > When fail occur i will setup new one zoo. But i have to restart all of solr > nodes? Or is there way to inform them that zoo has new ip, and make them > reconnect to it? > > > > -- > Other think I wanted to ask is how to replicate between two clasters. > Imagine that i got two clasters of solr in diferent zones. Lets say for > failover purpose. > > Having them all in one single claster covered by zoo will make appear leader > in zone that i probably dont want to have there. Can I somehow specify > order for nodes i want to appear as leader in case of ealier leader die? > > > > > > -- > View this message in context: > http://lucene.472066.n3.nabble.com/SolrCloud-new-zookeper-node-on-different-ip-replicate-between-two-clasters-tp4039101.html > Sent from the Solr - User mailing list archive at Nabble.com.
