I am thinking of having a cluster of one indexer and a few searchers 1 to n. The indexer will consist of a number of stages as defined in SEDA. I must still do this decomposition. the resulting index will be published via message q to the searchers that will stop doing searches long enough to update the local index. What is the purpose behind the nutch mapreduce component? Would it be useful for me to look at it in order to better decompose the indexer component? Cheers, Peter Gelderbloem
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]