Hi, We currently have a master-slave setup for solr with two slave servers. We are using Solrj (stream-update-solr-server) to index master slave, which takes 6 hours to index around 15 million documents.
I would like to explore hadoop, in particularly for indexing job using mapreduce approach. - I have read some comments on the JIRA tickets, but it still seems unclear how this setup will work. - I am not sure as what tasks will be done at map phase and what on reduce phase. - And would it merge the multiple indices together into one during reduce phase or is this a separate task out of mapreduce? Any directions and guidance over this setup would be highly appreciated. Thanks in advance, -Ali -- View this message in context: http://lucene.472066.n3.nabble.com/solr-with-hadoop-tp482688p914483.html Sent from the Solr - User mailing list archive at Nabble.com.
