Quick quesiton.
We are doing some tests over NFS right now with right at 1 million pages.
The first search takes a little over 1 second
Results Hits 1-10 (out of 101 total matching documents): for nutch [definition] (1.081 seconds)
The second time the results are right there for speed.
Results Hits 1-10 (out of 101 total matching documents): for nutch [definition] (0.063 seconds)
NFS will make searches slower. It is best to keep at least your indexes, and, better yet, segment data, on the local disks of search machines.
As a /segments grows bigger and bigger how can you ensure fast search results?
Limit the number of pages per search node by spreading your segments over a number of machines, using Nutch's distributed search code. The command 'bin/nutch server' starts a search server node. Then list the search hosts in a search-servers.txt file, one host/port pair per line. For best performance, each search node should have a merged index (as created by 'bin/nutch merge') searching all of the segments on that node.
Doug
------------------------------------------------------- SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33 Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift. http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285 _______________________________________________ Nutch-general mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/nutch-general
