Okay, so the matter is that when :

- I'm in multi-datacenters configuration (with endpoint_snitch: 
PropertyFileSnitch)
- I create a keyspace with strategy SimpleStrategy
- I create a column family in this keyspace and select from it with CONSISTENCY 
LOCAL_ONE

I get a rpc_timeout (A developer did all these actions :D ) I understand that 
SimpleStrategy can't work with multi-datacenters, but shouldn't we have an 
error raised when using this type of consistency ?

Another question is why can we still create SimpleStrategy keyspaces with 
PropertyFileSnitch enabled ?

-- 
Cyril SCETBON

On 14 Feb 2014, at 11:04, Cyril Scetbon <cyril.scet...@free.fr> wrote:

> Hi,
> 
> Good advice. I found earlier in the morning that it's related to consistency 
> LOCAL_ONE. I'll check later if it should raise an error in some cases.
> 
> Thanks
> -- 
> Cyril SCETBON
> 
> On 14 Feb 2014, at 10:12, Vivek Mishra <mishra.v...@gmail.com> wrote:
> 
>> Check for consisteny level and socket timeout setting on client side.
>> 
>> -Vivek
>> 
>> 
>> On Fri, Feb 14, 2014 at 2:36 PM, Cyril Scetbon <cyril.scet...@free.fr> wrote:
>> After a few tests, it does not depend on the query. Whatever cql3 query I 
>> do, I always get the same exception. If someone sees something ...
>> --
>> Cyril SCETBON
>> 
>> On 13 Feb 2014, at 17:22, Cyril Scetbon <cyril.scet...@free.fr> wrote:
>> 
>> > Hi,
>> >
>> > I get a weird issue with cassandra 1.2.13. As written in the subject, a 
>> > query executed by class CqlPagingRecordReader raises a TimedOutException 
>> > exception in Java but I don't have any error when I use it with cqlsh. 
>> > What's the difference between those 2 ways ? Does cqlsh bypass some 
>> > configuration compared to Java ?
>> >
>> > You can find my sample code at http://pastebin.com/vbAFyAys (don't take 
>> > care of the way it's coded cause it's just a sample code). FYI, I can't 
>> > reproduce it on another cluster. Here is the output of the 2 different 
>> > ways (java and cqlsh) I used http://pastebin.com/umMNXJRw
>> >
>> > Thanks
>> > --
>> > Cyril SCETBON
>> >
>> 
>> 
> 

Reply via email to