I'm not entirely sure but I think you can only use get_range_slices with start_key/end_key on a cluster using OrderPreservingPartitioner. Dont know if that is intentional or buggy like Jonathan suggest but I saw the same "duplicates" behaviour when trying to iterate all rows using RP and start_key/end_key.
The only solution I found was using start_token/end_token which also has some issues but once you work arround those everything works just fine (0.6.3 for me). On Tue, Jul 13, 2010 at 12:39 AM, Jonathan Ellis <[email protected]> wrote: > I can't picture how you could be reading data that sorts *before* the > start key with a range slice. So, probably not fixed in 0.6.3. > > On Mon, Jul 12, 2010 at 7:56 AM, Per Olesen <[email protected]> wrote: >> >>>This is a bug. Can you submit a ticket with test data to reproduce? >> >> Uuuh, maybe...:) >> >> Right now it is happening on some life user data, that I am not sure I can >> ship. Haven't tried if I can reproduce locally. >> >> One question: We are running 0.6.2. Could this be fixed in 0.6.3? Not that >> big a problem for us to upgrade, if you want me to check that first? >> >> /Per >> > > > > -- > Jonathan Ellis > Project Chair, Apache Cassandra > co-founder of Riptano, the source for professional Cassandra support > http://riptano.com >
