Update some replicator defaults Use smaller worker batch size so that overall progress is more steady. Also decrease connection pipeline size because an error in one request will result in all other subsequent requests in the same connection to be retried, making overall progress slower - this is more visible when doing multiple pull replications from the same server.
couchdb/trunk/etc/couchdb/default.ini.tpl.in
couchdb/trunk/src/couchdb/couch_replicator_utils.erl