Hi Olivier, the index size is relative big and you enabled replication after startup: <str name="replicateAfter">startup</str>
This could explain why the slave is replicating from the very beginning. Are the index versions/generations the same? (via command or admin/replication) If not, the slaves tries to replicate and if that fails it will try again and again. Are the servers in the same LAN? Maybe replication is just slow? BTW: CPU usage should be small while replication but peaks nearly after replication because of autowarming. Do you have enough free disc space on the slave? (>54GB) To understand what might be the problem comment out the two replicateAfter statements in the config + reload + go to the admin/replication site. There hit 'replicate now' and see what happens. Update the UI after some seconds. (not via F5 I think but via entering the url again + Enter to avoid POST this again) If nothing seems to change you'll have an error. Then look at the logs to investigate this and try the command on master and slave to better 'debug' replication: http://slave_host:port/solr/replication?command=details documented at: http://wiki.apache.org/solr/SolrReplication Hopes this helps a bit ... Regards, Peter. -- http://jetwick.com twitter search prototype