Hello, Sorry I would like to understand all components correctly and hava a question about DHT. Instead of loookup from all servers related file your are starting to use 32bit hash for each file for keep where is stored. 1) Where this hash is stored ? on servers or on clients ? I mean clients already know the hash and request file directly from related server ? or all backend server have all hash informations and redirect request to right server ? 2) ?n where and which format are you storing hashs , can I see it ? 3) Looks like if I add new server to servers farm then looks like no way to distribute already stored data to new server for balance disk usage on all servers ? Because file is already hashed and there is no rehesh mechanisim after data rearranged, right ? if it is , then design problem could couse because if people do not know trend of their data, they can do wrong things and they can not optimize it in feaure , isn't it ? Regards Vahric -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://zresearch.com/pipermail/gluster-users/attachments/20090530/e0b71c58/attachment.htm>
